WO2011021376A1

WO2011021376A1 - Control device and control method for robot arm, household robot, control program for robot arm, and integrated electronic circuit for controlling robot arm

Info

Publication number: WO2011021376A1
Application number: PCT/JP2010/005055
Authority: WO
Inventors: 優子津坂; 安直岡▲崎▼
Original assignee: パナソニック株式会社
Priority date: 2009-08-21
Filing date: 2010-08-12
Publication date: 2011-02-24

Abstract

A control device for a robot arm (5) for performing housework at home is provided with a motion database (17) in which information relating to the motion of the robot arm (5) is recorded, a correction motion type determination unit (23) which determines the type of correction of the motion, a force detection unit (53) which detects the force of a person (16), an object state identification unit (19) which identifies the state of an object, and a motion correction unit (20) which corrects the motion according to the force of the person (16) and the type of correction while the robot arm (5) is working.

Description

Robot arm control device and control method, housework robot, robot arm control program, and integrated electronic circuit for robot arm control

The present invention relates to a robot arm control device and control method for generating and teaching a housework operation of a robot that performs housework in a home, a robot arm and a housework robot having the control device, a robot arm control program, and a robot arm control The present invention relates to an integrated electronic circuit.

In recent years, home robots such as care robots or housework support robots have been actively developed. Unlike a robot for industrial use, a home robot is operated by an amateur at home, so it is necessary to be able to easily teach the operation. Furthermore, since the operating environment when the robot is working varies depending on the home, it is necessary to flexibly cope with the home environment.

As an example of the teaching method of the robot apparatus, a force sensor is attached to the wrist of the robot, the teaching operator directly holds the handle attached to the tip of the force sensor, and the robot is guided to the teaching point. Teaching is performed (see Patent Document 1).

Furthermore, when the teaching worker teaches the robot by directly gripping the robot, the teaching worker understands the intention of the teaching worker and automatically changes the operational feeling of force control during the teaching operation (patent) Reference 2).

JP 59-157715 A JP 2008-110406 A

However, in Patent Document 1, since it is necessary for the teaching worker to teach all teaching points, teaching takes time and is very troublesome. Furthermore, in the industrial field, when correcting a part of the taught movement, it must be corrected by programming with a remote device called a teaching pendant, or all operations must be taught from the beginning. Was bad.

Furthermore, in Patent Document 2, the teaching worker understands the intention of the teaching worker and directly changes the operational feeling during the operation when the teaching worker directly teaches. It does not understand the operation intention other than the operational feeling such as which parameter is to be operated among a plurality of types of teaching parameters such as speed. Therefore, it is necessary for the work teacher to explicitly set which parameter the work teacher teaches. Furthermore, part of the taught movement could not be corrected, and work efficiency was poor.

An object of the present invention has been made in view of such problems, and a robot arm control device and method, a housework robot, and a robot arm that enable an operator to teach a robot easily and in a short time. And an integrated electronic circuit for controlling a robot arm.

In order to achieve the above object, the present invention is configured as follows.

According to the first aspect of the present invention, there is provided a control device for a robot arm that controls the operation of the robot arm and performs housework work while acting on the object of housework work by the robot arm in the home,
Force detecting means for detecting the force of a person acting on the robot arm;
An information acquisition unit for acquiring information on the operation including the position of the robot arm in the housework and information on the human force detected by the force detection unit;
Object state determination means for determining the state of the object;
The motion from the information regarding the motion including the position of the robot arm in the housework work acquired by the information acquisition unit, the information regarding the force of the person, and the state of the target determined by the target state determination means. Corrective action type determining means for determining the type of corrective action for correcting
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the domestic work of the robot arm determined in advance. There is provided a control device for a robot arm, comprising a motion correction means for controlling the robot arm and correcting the motion according to the type.

According to a fifteenth aspect of the present invention, there is provided a control method for a robot arm that controls the operation of the robot arm and performs housework work while acting on an object of housework work by the robot arm in the home,
Detecting the force of the person acting on the robot arm with force detection means,
Information on the operation including the position of the robot arm in the housework and information on the human force detected by the force detection means are respectively acquired by the information acquisition unit,
The state of the object is determined by the object state determination means,
The motion from the information regarding the motion including the position of the robot arm in the housework work acquired by the information acquisition unit, the information regarding the force of the person, and the state of the target determined by the target status determination means. The corrective action type determining means determines the type of corrective action to correct the human force detected by the force detecting means and acquired by the information acquisition unit during the housework of the robot arm determined in advance. A control method of the robot arm, wherein the robot arm is controlled and the motion is corrected by the motion correction means according to the information relating to the information and the type of the correction motion determined by the correction motion type determination means provide.

According to a sixteenth aspect of the present invention, the robot arm;
There is provided a domestic robot comprising the robot arm control device according to any one of the first to fourteenth aspects for controlling the robot arm.

According to a seventeenth aspect of the present invention, there is provided a control program for a robot arm that controls the operation of the robot arm and performs housework work while acting on the object of housework work by the robot arm in the home,
Information related to the operation including the position of the robot arm in the housework and information detected by the force detection means and acquired by the information acquisition unit and determined by the object state determination means Determining a type of correction operation for correcting the operation from the state of the object by a correction operation type determination unit;
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the housework work of the robot arm determined in advance and the correction operation type determination unit There is provided a robot arm control program for causing a computer to execute an operation correction step of controlling the robot arm and correcting the operation by an operation correction unit according to the type.

According to an eighteenth aspect of the present invention, there is provided an integrated electronic circuit for controlling a robot arm that controls the operation of a robot arm and performs domestic work while acting on an object of domestic work by the robot arm in the home,
Information related to the operation including the position of the robot arm in the housework and information detected by the force detection means and acquired by the information acquisition unit and determined by the object state determination means A correction operation type determining means for determining a type of correction operation for correcting the operation from the state of the object;
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the housework work of the robot arm determined in advance and the correction operation type determination unit There is provided an integrated electronic circuit for controlling a robot arm, comprising: an operation correcting unit that controls the robot arm according to the type to correct the operation.

As described above, according to the robot arm control device and the housework robot of the present invention, the force detection means, the information acquisition unit, the object state determination means, the correction action type determination means, and the action correction means. By having the robot, the robot can easily correct the housework operation according to the force of the person using the information about the action including the position of the robot arm in the housework and the information about the person's force. The arm can be controlled.

According to the robot arm control method, robot arm control program, and robot arm control integrated electronic circuit of the present invention, the robot arm control integrated electronic circuit includes the correction operation type determination means and the operation correction means. It is possible to control the robot arm that can easily correct the housework operation according to the information related to the operation including the position of the robot arm and the information related to the human force.

Furthermore, by providing the correction operation type determination means, it is possible to automatically switch and correct a plurality of operations without using a button or the like.

Furthermore, by having the correction operation type determination means, it is possible to switch between performing a plurality of types of corrections at a time or performing one type of correction in accordance with the skill of the operator.

Further, by further comprising the control parameter management means and the impedance control means, by setting the mechanical impedance value of the robot arm according to the type of the correction operation, the mechanical impedance according to the correction direction of the robot arm. It is possible to control by changing the value, or to weaken or stop the suction force or the force on the work surface being corrected.

These and other objects and features of the invention will become apparent from the following description taken in conjunction with the preferred embodiments with reference to the accompanying drawings. In this drawing,
FIG. 1 is a diagram showing an outline of a configuration of a control device for a robot arm in the first embodiment of the present invention. FIG. 2A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 2B is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 3 is a diagram showing a detailed configuration of the robot arm control device and the robot arm that is a control target that constitute the housework robot in the first embodiment of the present invention; FIG. 4 is a diagram illustrating a list of operation information in an operation database of the control device for the robot arm in the first embodiment of the present invention; FIG. 5 is a diagram for explaining information about an operation database flag of the robot arm control device according to the first embodiment of the present invention; FIG. 6 is a diagram for explaining information relating to a correction parameter flag of the robot arm control device according to the first embodiment of the present invention; FIG. 7 is a block diagram illustrating a configuration of a control unit of the robot arm control device according to the first embodiment of the present invention; FIG. 8 is a diagram relating to the path of the control device for the robot arm in the first embodiment of the present invention, FIG. 9 is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 10 is a diagram for explaining a list of work impossible area database information of the robot arm control device according to the first embodiment of the present invention; FIG. 11 is a diagram relating to a path of the robot arm control device according to the first embodiment of the present invention; FIG. 12A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 12B is a plan view (a view of FIG. 12A seen from above) showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 12C is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 12D is a plan view showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 13A is a diagram relating to a coordinate system of the control device for the robot arm in the first embodiment of the present invention; FIG. 13B is a diagram related to a coordinate system of the control device for the robot arm in the first embodiment of the present invention; FIG. 13C is a diagram related to a coordinate system of the control device for the robot arm in the first embodiment of the present invention; FIG. 14 is a flowchart showing the operation steps of the correction operation type determination unit of the control device for the robot arm in the first embodiment of the present invention; FIG. 15 is a diagram illustrating a relationship between the force applied by a person of the control device for the robot arm and the time thereof in the first embodiment of the present invention; FIG. 16A is a side view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 16B is a plan view showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 16C is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 17 is a flowchart showing the operation steps of the correction operation type determination unit of the control device for the robot arm in the first embodiment of the present invention; FIG. 18A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 18B is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 19 is a diagram illustrating a correspondence between a force applied by a person of the control device for the robot arm and a suction force in the first embodiment of the present invention; FIG. 20A is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 20B is an enlarged plan view of the suction nozzle when explaining the operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 21 is a diagram illustrating a display unit of a peripheral device of the control device for the robot arm in the first embodiment of the present invention; FIG. 22 is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 23 is a diagram illustrating an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 24 shows operation steps of an operation correction unit, a correction operation type determination unit, an operation selection unit, an operation storage unit, an operation database, and a control parameter management unit of the robot arm control device according to the first embodiment of the present invention. It is a flowchart, FIG. 25 is a flowchart showing the operation steps of the control unit of the robot arm control device according to the first embodiment of the present invention; FIG. 26 is a diagram illustrating a data input IF of a peripheral device of the robot arm control device according to the first embodiment of the present invention. FIG. 27A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 27B is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 27C is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 28A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 28B is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 28C is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 29A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 29B is a plan view (a view of FIG. 29A seen from above) showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 29C is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 29D is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 30A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 30B is a plan view (a view of FIG. 30A seen from above) showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 30C is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 30D is a plan view showing an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 31 is a diagram illustrating an operation state of the robot arm control device according to the first embodiment of the present invention; FIG. 32A is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 32B is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 32C is a diagram showing an operation state of the control device for the robot arm in the first embodiment of the present invention; FIG. 33 is a diagram for explaining the force and position threshold values of the robot arm control device according to the first embodiment of the present invention; FIG. 34 is a diagram showing an outline of the configuration of the control device for the robot arm in the second embodiment of the present invention. FIG. 35 is a diagram for explaining a list of operation information in the operation database of the control device for the robot arm in the second embodiment of the present invention; FIG. 36A is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 36B is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 36C is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 37 is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention, FIG. 38A is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 38B is a plan view showing the operation state of the control device for the robot arm according to the second embodiment of the present invention (a view of FIG. 38A as viewed from above); FIG. 38C is a plan view showing an operation state of the robot arm controller in the second embodiment of the present invention; FIG. 38D is a plan view showing an operation state of the robot arm control device according to the second embodiment of the present invention; FIG. 39A is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 39B is a diagram illustrating an operation state of the robot arm control device according to the second embodiment of the present invention; FIG. 39C is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 40A is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 40B is a diagram illustrating an operation state of the robot arm control device according to the second embodiment of the present invention; FIG. 40C is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 41A is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 41B is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 41C is a diagram showing an operation state of the control device for the robot arm in the second embodiment of the present invention; FIG. 42 is a flowchart showing the operation steps of the correction operation type determination unit of the control device for the robot arm in the second embodiment of the present invention; FIG. 43 is a diagram illustrating a detailed configuration of the robot arm control device and the robot arm that is a control target that configure the housework robot according to the second embodiment of the present invention; FIG. 44A is a diagram illustrating a list of work procedure information in a work procedure information database of the control device for the robot arm in the second embodiment of the present invention; FIG. 44B is a diagram illustrating a list of work procedure information in a work procedure information database of the control device for the robot arm in the second embodiment of the present invention; FIG. 44C is a diagram illustrating a list of work procedure information in the work procedure information database of the robot arm control device according to the second embodiment of the present invention; FIG. 45 is a diagram illustrating a detailed configuration of the robot arm control device and the robot arm that is a control target that constitute the housework robot according to the second embodiment of the present invention.

Before continuing the description of the present invention, the same parts are denoted by the same reference numerals in the accompanying drawings.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

Hereinafter, various embodiments of the present invention will be described before the embodiments of the present invention are described in detail with reference to the drawings.

構成 With this configuration, the housework movement of the robot arm can be corrected according to human power.

According to the second aspect of the present invention, the information related to the operation includes information on the position of the robot arm according to the housework performed by the robot arm, and force information applied to the work surface from the robot arm, Among the information regarding the direction of the robot arm, the information regarding the strength of the suction force of the robot arm, the speed information of the robot arm, and the work disabling area information which is information regarding the area where the robot arm is not operated The robot arm control device according to the first aspect is characterized by having at least one of the following information.

With such a configuration, the position information, the force information applied by the robot arm, the information on the direction, the information on the strength of suction, and the speed information at each time according to the work performed by the robot arm. And at least one piece of information related to an area that you do not want to work on can be corrected.

According to a third aspect of the present invention, the information related to the movement includes information on force applied from the robot arm to the work surface according to the housework performed by the robot arm, and strength of the suction force of the robot arm. And at least information about
The motion correction unit is configured to apply a force control mode in which a predetermined force is applied from the robot arm to the work surface based on the information related to the motion in the xyz axis direction in which the robot arm can move. The operation before the correction operation according to the force of the person detected by the force detection means and acquired by the information acquisition unit while performing the operation by the robot arm by setting for each axis The robot arm control device according to the first aspect is characterized in that the magnitude or direction of the set force among the information regarding the correction is corrected.

With such a configuration, a force control mode for performing the operation by applying a preset force from the robot arm to the work surface based on the information related to the operation in the xyz-axis direction in which the robot arm can move. While the operation is being performed by the robot arm by setting for each axis, the setting of the information related to the operation before the correction operation is set according to the force of the person detected by the force detection means. The magnitude or direction of the applied force can be corrected.

According to a fourth aspect of the present invention, the information related to the movement includes information on the position of the robot arm, information on the direction of the robot arm, and the robot arm according to the housework performed by the robot arm. Speed information and work non-working area information that is information related to a work non-working area,
The motion correction means is configured to apply a force applied to the robot arm from the person to the robot arm while operating in a position control mode for controlling the position of the robot arm based on information on the motion. In response, the impedance detection mode in which the robot arm operates is set for each axis in the xyz axis direction in which the robot arm can move, and the operation is performed while the operation is being performed. The robot arm control device according to the first aspect is characterized in that the operation of the information related to the operation in the impedance control is corrected according to the human force acquired by the information acquisition unit.

With such a configuration, from the person to the robot arm that has stopped driving while operating in the position control mode for controlling the position of the robot arm based on the information related to the operation, The force detection unit is configured to operate the operation while setting an impedance control mode in which the robot arm operates according to a force applied to the arm for each axis in the xyz axis direction in which the robot arm can move. The operation of the information related to the operation in the impedance control can be corrected according to the force of the person detected in (1).

According to the fifth aspect of the present invention, control parameter management means for setting a mechanical impedance setting value of the robot arm based on the type of the correction action determined by the correction action type determination means;
In any one of the first to fourth aspects, the apparatus further comprises impedance control means for controlling the mechanical impedance value of the robot arm to the mechanical impedance set value set by the control parameter management means. A control device for the described robot arm is provided.

With such a configuration, it is possible to set and control the mechanical impedance value of the robot arm based on the type of correction operation.

According to the sixth aspect of the present invention, the impedance control means individually sets the mechanical impedance setting values in the six axes of the translational direction and the rotational direction of the hand of the robot arm based on the type of the correction operation. And
Further, when correcting the direction of the robot arm of the hand as the type of the correction operation determined by the correction operation type determination unit, the control parameter management means, the rigidity of the robot arm in the direction to be corrected, The robot arm control device according to the fifth aspect, wherein the mechanical impedance set value is set so as to be higher than the rigidity of the robot arm in a direction different from the direction to be corrected.

With such a configuration, as the type of correction, the direction in which the hand is desired to be corrected is made highly rigid, so that it becomes easier to detect the amount of change in the direction to be corrected and the robot arm can be moved in the direction to be corrected. By making the direction different from the direction to be corrected low and highly rigid, it can be made difficult to move.

According to the seventh aspect of the present invention, the correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
When the force component in the direction perpendicular to the work surface is equal to or less than a first threshold and the force component in the direction parallel to the work surface is equal to or greater than a second threshold, When the amount of movement of the position of the hand of the robot arm detected by the correction operation type determining means in the direction parallel to the work surface is equal to or greater than a third threshold value, the position of the work surface is determined as the type of correction operation. Decide that it is the type of movement,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to the first aspect, wherein the position of the hand of the robot arm is corrected in a direction parallel to the work surface.

According to the eighth aspect of the present invention, the correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is greater than a fourth threshold, the type of the correction operation is determined as the type of movement of the position in the direction perpendicular to the work surface,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to the first aspect, wherein the position of the hand of the robot arm is corrected in a direction perpendicular to the work surface.

According to the ninth aspect of the present invention, the correction operation type determining means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is less than or equal to a fourth threshold and the housework operation is a wiping operation, the type of correction operation is determined to be a correction type of force application ,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination means, The robot arm control device according to the first aspect, wherein the force applied by the robot arm in a direction perpendicular to the work surface is corrected.

According to the tenth aspect of the present invention, the correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is a fourth threshold value or less and the housework work is a suction work, it is determined that the type of the correction operation is a type of correction of the suction force,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to the first aspect, wherein the suction force in a direction perpendicular to the work surface is corrected.

According to the eleventh aspect of the present invention, the correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
The correction operation type determination unit is configured to correct the correction when a force component in a direction perpendicular to the work surface is less than a first threshold value and a force component in a direction parallel to the work surface is greater than or equal to a second threshold value. When the movement amount in the direction parallel to the work surface of the position of the hand of the robot arm detected by the action type determining means is less than a third threshold, the correction action type is a speed correction type. Determined that there was
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to the first aspect, wherein the speed of the position of the hand of the robot arm is corrected in a direction parallel to the work surface.

According to a twelfth aspect of the present invention, the correction action type determination means is a force applied to the robot arm based on the force of the person applied to the robot arm detected by the force detection means and acquired by the information acquisition unit. If the displacement amount of the posture is larger than the displacement amount of the position component, the type of the correction operation is determined. Determined to be the type of posture correction,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination means, The robot arm control device according to the first aspect, wherein the posture of the hand of the robot arm is corrected.

With such a configuration, according to the human force detected by the force detection unit and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination unit, the robot arm The drive device of the robot arm can be reliably driven and controlled so as to correct the posture of the hand.

According to the thirteenth aspect of the present invention, the correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
The correction operation type determination means is configured so that a force applied to the robot arm by the human hand is parallel to the work surface and parallel to the work surface for a certain time detected by the correction operation type determination means. When the amount of movement in the direction is equal to or greater than a threshold, it is determined that the type of correction operation is the type of setting of the work impossible area,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to the first aspect, wherein the work impossible area is set by moving the position of the hand of the robot arm.

With such a configuration, the work-impossible area can be easily set, and work can be made unnecessary for an area that is not desired to be worked.

According to a fourteenth aspect of the present invention, the apparatus further comprises display means for displaying information on the type of the correction operation based on the type of the correction operation determined by the correction operation type determination unit. A robot arm control device according to any one of the thirteenth aspects is provided.

With this configuration, it is possible to display information regarding the type of correction operation.

According to a fifteenth aspect of the present invention, there is provided a control method for a robot arm that controls the operation of the robot arm and performs housework work while acting on an object of housework work by the robot arm in the home,
Detecting the force of the person acting on the robot arm with force detection means,
Information on the operation including the position of the robot arm in the housework and information on the human force detected by the force detection means are respectively acquired by the information acquisition unit,
The state of the object is determined by the object state determination means,
The motion from the information regarding the motion including the position of the robot arm in the housework work acquired by the information acquisition unit, the information regarding the force of the person, and the state of the target determined by the target state determination means. The type of correction operation for correcting the correction is determined by the correction operation type determination means,
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the domestic work of the robot arm determined in advance. A robot arm control method is provided, wherein the robot arm is controlled according to the type, and the motion is corrected by a motion correction means.

With such a configuration, the correction type of the movement is determined based on information on the movement of the robot arm, the human force is detected, and the work of the robot arm is performed according to the human force and the correction type. The operation can be corrected.

According to a seventeenth aspect of the present invention, there is provided a control program for a robot arm that controls the operation of the robot arm and performs housework work while acting on the object of housework work by the robot arm in the home,
Information on the operation including the position of the robot arm in the housework and information on the force of the person acting on the robot arm detected by the force detection unit and acquired by the information acquisition unit and determined by the object state determination unit Determining a type of correction operation for correcting the operation from the state of the object by a correction operation type determination means;
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the housework work of the robot arm determined in advance and the correction operation type determination unit There is provided a robot arm control program for causing a computer to execute an operation correction step of controlling the robot arm and correcting the operation by an operation correction unit according to the type.

According to an eighteenth aspect of the present invention, there is provided an integrated electronic circuit for controlling a robot arm that controls the operation of a robot arm and performs domestic work while acting on an object of domestic work by the robot arm in the home,
Information on the motion including the position of the robot arm in the housework and information on the force of the person acting on the robot arm detected by the force detection means and acquired by the information acquisition unit and determined by the object state determination means Corrective action type determining means for determining the type of corrective action for correcting the action from the state of the object;
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the housework work of the robot arm determined in advance and the correction operation type determination unit There is provided an integrated electronic circuit for controlling a robot arm, comprising: an operation correcting unit that controls the robot arm according to the type to correct the operation.

(First embodiment)
First, the configuration of the household robot 1 according to the first embodiment of the present invention will be described. As shown in FIG. 1, an explanation will be given by taking as an example a case of performing an operation of polishing equipment 6 such as an IH cooking heater or a kitchen workbench or a suction operation of dust (in other words, a cleaning operation) in the home. To do.

The household robot 1 includes a robot arm 5 and a control device 1000 that drives and controls the robot arm 5, and drives and controls the driving

devices

65, 67, 43, and 62 with the control device 1000 in the home. The above work is performed.

The robot arm 5 is installed on a wall surface 7a of a work table 7 such as a kitchen or a table in the home, and the base end of the robot arm 5 is movably supported on a rail 8 fixed to the wall surface 7a. The robot arm 5 can move in the lateral direction along the rail 8, for example, in the horizontal direction, automatically by the force of the hand 16 of the person 16 </ b> A or by a motor or the like. The fixing position of the base end of the robot arm 5 is not limited to the wall surface 7a of the work table 7, but may be a ceiling or the like.

The side surface of the work table 7 includes a data input IF 26 such as an operation panel 26A on which buttons 26a and the like are arranged, and a display unit 14 as an example of display means arranged on the wall surface of the work table 7 and the like. Has been.

The rail 8 includes a rail fixing portion 8a fixed to the wall surface 7a, and a wheel (not shown) that is driven to rotate forward and backward by driving a motor 65 that is an example of a driving device for the rail movable portion. It is comprised with the rail movable part 8b which can move with respect to 8a. The base part 34 to which the base end of the robot arm 5 is connected is connected to the rail movable part 8b, and the base part 34 of the robot arm 5 is configured to be movable together with the rail movable part 8b with respect to the rail fixing part 8a. ing. Alternatively, instead of such a configuration, a wheel that is driven to rotate forward and backward by a motor 65 is provided on the base portion 34 to which the base end of the robot arm 5 is connected, and along the rail 8 fixed to the wall surface 7a, It is good also as a structure that the base part 34 moves.

A suction nozzle 9 as an example of a cleaning unit is detachably attached to the tip of the robot arm 5 so as to be detachable. The suction nozzle 9 is rotatably accommodated in the suction nozzle 9 and is rotationally driven by a rotary brush motor 67 which is an example of a rotary brush drive device in the suction nozzle 9, so that the work surface (cleaning surface) of the device 6 is driven. A rotating brush 11 that sweeps up dust is disposed. Reference numeral 10 denotes a mop as another example of a cleaning unit that can be detachably attached to the tip of the robot arm 5 in place of the suction nozzle 9 to wipe off the dirt of the device 6.

30 is a hand arranged at the tip of the robot arm 5 and, for example, functions as a mechanism for detachably attaching the suction nozzle 9 and replacing the suction nozzle 9 with a mop 10 of another cleaning unit. As such a mechanism, for example, a known detachable attachment mechanism such as a mechanical chuck can be used.

The housework robot 1 is a robot that performs a work of sucking dust or dust of the device 6 with the suction nozzle 9, a work of wiping the dirt of the device 6 with the mop 10, or a work of polishing with the mop 10 with force.

The outline of the operation procedure of the household robot 1 will be described.

First, in FIG. 2A, the hand 16 of the person 16A is turned on by the data input IF 26 (for example, pressing “ON” of the power button 26a of the operation panel 26A of FIG. 26) arranged on the side surface of the work table 7.

Next, when sucking dust or dust or the like, the suction nozzle 9 is attached to the hand 30 at the tip of the robot arm 5 of the house robot 1 with a human hand 16, and when cleaning or polishing is performed, the house robot 1 The mop 10 is attached to the hand 30 at the tip of the robot arm 5 with the hand 16 of the person 16A. When attaching the suction nozzle 9 or the mop 10 with the hand 16 of the person 16A, data input from the data input IF 26 such as a button (for example, “open / close button 26b of the operation panel 26A of FIG. 26 for opening / closing the hand 30” An instruction to open the hand 30 is issued to the control unit 22 (to be described later) of the household robot 1 to open the hand 30 by pressing “open” or the like. After that, the suction nozzle 9 or the mop 10 is attached to the hand 30 and data is input from the data input IF 26 (for example, pressing the “close” button of the open / close button 26b for opening / closing the hand 30 on the operation panel 26A in FIG. 26). The suction nozzle 9 or the mop 10 is attached to the hand 30 by issuing an instruction to the control unit 22 to close the hand 30 and closing the hand 30. At the time of attachment, the tip of the robot arm 5 is moved (for example, by pressing “open” of the open / close button 26b, the hand 30 at the tip of the robot arm 5 is automatically raised to the upward position. In order to facilitate the operation of the hand 30, it can be operated so as to come to the hand 16 of the person as shown in FIG. 2B. For example, the hand 30 at the tip of the robot arm 5 can be automatically lowered to the downward cleaning position by pressing “close” of the open / close button 26b.

Next, when the data input IF 26 (for example, the start button of the cleaning switch 26c of the operation panel 26A in FIG. 26) disposed on the side surface of the workbench 7 is pushed with the hand 16 of the person 16A, the house robot 1 is activated. Then, an optimum operation, for example, a cleaning operation (for example, suction or wiping operation) is selected by an operation selection unit 29 to be described later, and a cleaning operation (for example, suction or wiping operation) based on the selected cleaning operation is started.

During the wiping and cleaning operation, as shown in FIGS. 16A and 16B (views of FIG. 16A from above), the domestic robot 1 self-runs on the rail 8, and simultaneously with the self-running operation, the tip of the robot arm 5 The mop 10 performs a wiping and cleaning operation with a trajectory that draws a spiral, for example, while gradually shifting in the left-right direction around the position along the central axis along the front-rear direction of the robot arm 5. In the suction cleaning operation, as shown in FIG. 16C, the housework robot 1 self-runs on the rail 8, and simultaneously with the self-running operation, the robot arm 5 is driven to suck the suction nozzle at the tip of the robot arm 5. 9 performs suction cleaning by moving in a direction perpendicular to the self-running direction (that is, reciprocating along the front-rear direction perpendicular to the left-right direction).

The operation panel 26A as an example of the data input IF 26 is fixed to the side surface of the work table 7. However, a remote control capable of remote operation may be used.

Next, the person confirms the degree of dirt on the work surface, directly grips the robot arm 5 of the household robot 1 with the hand 16 of the person 16A, and corrects the cleaning operation (for example, toward a region where the degree of dirt is severe). For example, as shown in FIG. 12A, the robot arm 5 or the housework robot of the housework robot 1 is applied by applying a force to the suction nozzle 9 or the mop 10 at the tip of the robot arm 5. 1 operation is corrected. That is, as shown in FIG. 12C, when the suction nozzle 9 or the mop 10 is moved as shown by a solid line in the zigzag direction, for example, as shown in FIG. The tip of the robot arm 5 or the suction nozzle 9 or the mop 10 is moved by a human hand 16 as shown by the arrow so as to move in the direction indicated by the dotted line, and the suction nozzle at the tip of the robot arm 5 is moved. 9 or mop 10 is moved to the left. In this way, as shown in FIG. 12D, the suction nozzle 9 or the mop 10 is located in the left region with respect to the position along the central axis along the front-rear direction of the hand of the robot arm 5. The cleaning operation by the mop 10 can be performed as indicated by a solid line in the zigzag direction, for example.

FIG. 3 is a diagram showing in detail the components of the control device of the robot arm 5 constituting the housework robot 1, and includes a control device main body 45, a motion generation device 12 that generates a motion, and a robot arm that is a control target. 5 is a diagram showing a detailed configuration of the rail 5, the rail 8, and the peripheral device 47. FIG. The control device of the housework robot 1 is roughly composed of a control device main body 45, the motion generation device 12, and a peripheral device 47.

The control device main body 45, the motion generation device 12, and the peripheral device 47 are each configured by a general personal computer.

The control device main body 45 includes an operation correction unit 20 as an example of an operation correction unit of the operation generation device 12, a correction operation type determination unit 23 as an example of a correction operation type determination unit, and a data input IF 26 of a peripheral device 47, respectively. Control parameter management unit 21 as an example of connected control parameter management means, and control unit (impedance control unit) as an example of impedance control means connected to control parameter management unit 21 and input / output IF 24 of peripheral device 47 22.

The motion generation device 12 includes a motion database 17, an unworkable area database 28, a corrected motion type determination method setting unit 27, a motion correction unit 20, a corrected motion type determination unit 23, a motion storage unit 15, and a motion selection. It is comprised so that the part 29 and the information acquisition part 100 may be provided. The motion storage unit 15 is connected to the motion database 17, the work impossible area database 28, and the motion correction unit 20. The motion database 17 and the work impossible area database 28 are connected to the motion storage unit 15, the motion correction unit 20, and the motion selection unit 29, respectively. The motion correction unit 20 is connected to the motion database 17, the work impossible region database 28, the motion storage unit 15, the control parameter management unit 21 of the control device main body 45, the correction motion type determination unit 23, and the data input IF 26 of the peripheral device 47. Has been. The correction operation type determination unit 23 is connected to the operation correction unit 20, the correction operation type determination method setting unit 27, the data input IF 26 of the peripheral device 47, and the control parameter management unit 21 of the control device main body 45. The action selection unit 29 is connected to the action database 17, the work impossible area database 28, and the data input IF 26. The correction operation type determination method setting unit 27 is connected to the data input IF 26 and the correction operation type determination unit 23. The information acquisition unit 100 is connected to the correction operation type determination unit 23, the operation database 17, the work impossible area database 28, and the force detection unit 53 of the control unit 22. Therefore, the information acquisition unit 100 acts on the information regarding the cleaning operation including the suction force of the

cleaning units

9 and 10 and the cleaning position of the

cleaning units

9 and 10 in the cleaning operation, and the robot arm 5 detected by the force detection unit 53. It is possible to obtain information on human power. The information acquired by the information acquisition unit 100 is input to the correction operation type determination unit 23, and the cleaning operation is corrected as described later from the information regarding the cleaning operation and the information regarding human power acquired by the information acquisition unit 100, respectively. The type of corrective action to be performed can be determined by the corrective action type determining means 23.

The peripheral device 47 includes a correction operation type determination unit 23, an operation correction unit 20, a control parameter management unit 21 of the control device body unit 45, a data input IF 26 connected to the display unit 14 and the operation generation device 12, and a rail movable unit. An encoder 64 that is attached to the rotation shaft of the motor 65 of 8b and detects the rotation angle of the rotation shaft, and a rotation shaft of the motor 43 that is an example of a joint drive device for each joint portion. An encoder 44 for detecting an angle and an encoder 68 for detecting a rotation angle of the rotating shaft 11 attached to a rotating shaft of a hand opening / closing driving motor 62 which is an example of a hand driving device and an encoder 68 of a motor 69 of the rotating brush 11. And the input / output IF 24 connected to the control unit 22, the motor 65 of the rail movable unit 8b, A motor driver 25 connected to the motor 43 of each joint part of the bot arm 5, the motor 62 for driving the hand opening / closing, and the motor 69 of the rotary brush 11, and the display unit 14 connected to the correction operation type determination unit 23; It is comprised so that it may comprise.

The input / output IF 24 is configured to include, for example, a D / A board, an A / D board, a counter board, and the like connected to an expansion slot such as a PCI bus of a personal computer.

The motion generation device 12 that controls the motion of the robot arm 5 and the rail movable portion 8b, the control device main body 45, and the peripheral device 47 execute the respective motions, whereby each of the joint portions of the robot arm 5 is performed. Each joint angle information that is joint angle information and output from an encoder 44 described later is taken into the control device main body 45 through the input / output IF 24. Then, the control device main body 45 calculates a control command value in the rotation operation of each joint of the robot arm 5 based on each taken joint angle information. Further, the position information of the rail movable portion 8 b output from the encoder 64 of the motor 65 of the rail movable portion 8 b is taken into the control device main body 45 through the input / output IF 24. And the control apparatus main-body part 45 calculates the control command value of the motor 65 of the rail movable part 8b based on each taken-in positional information. Further, the rotational force output from the encoder 66 of the motor 67 of the rotary brush 11 is taken into the control device main body 45 through the input / output IF 24, and the control device main body 45 receives the rotational force of the rotary brush 11 based on the taken rotational force. A control command value for the motor 67 is calculated.

The calculated control command value of the motor 43 of each joint part of the robot arm 5 is given to the motor driver 25 through the input / output IF 24, and each joint part of the robot arm 5 according to each control command value sent from the motor driver 25. The motors 43 are driven independently.

Further, the calculated control command value of the rail movable portion 8b8 is given to the motor driver 25 through the input / output IF 24, and the motor 65 of the rail movable portion 8b is driven according to each control command value sent from the motor driver 25. .

Further, as an example of a hand drive device that is driven and controlled by the motor driver 25, a hand opening / closing drive motor 62 and an encoder 61 that detects the rotational phase angle of the rotation shaft of the hand opening / closing drive motor 62 are further provided in the hand 30. For example, the hand 30 is opened by rotating the rotating shaft of the motor 62 in the forward direction so that the suction nozzle 9 or the mop 10 can be attached by a human hand 16, while the rotating shaft of the motor 62 is reversed. The hand 30 can be closed by rotating in the direction, and the suction nozzle 9 or the mop 10 attached to the hand 30 can be fixed. In such a case, based on the rotation angle of the rotation shaft of the motor 62 detected by the encoder 61, a control signal (open / close command) from the hand control unit 54 (shown in FIG. 7) of the control unit 22 of the control device main body 45. Signal), the rotation of the hand opening / closing driving motor 62 is driven and controlled via the motor driver 25, and the hand 30 is opened / closed by rotating the rotating shaft of the hand opening / closing driving motor 62 forward and backward.

The calculated control command value of the motor 67 of the rotary brush 11 is given to the motor driver 25 through the input / output IF 24, and the motor 69 of the rotary brush 11 is driven according to the control command value sent from the motor driver 25. .

The robot arm 5 is a multi-link manipulator with 6 degrees of freedom, and the tip of the hand 30, the forearm link 32 having the wrist 31 to which the hand 30 is attached at the tip, and the proximal end of the forearm link 32 are rotatable. The upper arm link 33 to be connected is provided, and a base portion 34 on which the base end of the upper arm link 33 is rotatably connected and supported. The base part 34 is connected to the rail movable part 8b. The wrist portion 31 has three rotation axes of the fourth joint portion 38, the fifth joint portion 39, and the sixth joint portion 40, and changes the relative posture of the hand 30 with respect to the forearm link 32. be able to. That is, in FIG. 3, the fourth joint portion 38 can change the relative posture around the horizontal axis of the hand 30 with respect to the wrist portion 31. The sixth joint portion 40 can change the relative posture of the hand 30 with respect to the wrist portion 31 about the horizontal axis orthogonal to the horizontal axis of the fourth joint portion 38 and the vertical axis of the fifth joint portion 39. . The other end of the forearm link 32 is rotatable around the third joint portion 37 with respect to the tip of the upper arm link 33, that is, around a horizontal axis parallel to the horizontal axis of the fourth joint portion 38. The other end of the upper arm link 33 is rotatable around the second joint portion 36 with respect to the base portion 34, that is, around a horizontal axis parallel to the horizontal axis of the fourth joint portion 38. Further, the upper movable portion 34a of the pedestal 34 rotates around the first joint 35 relative to the lower fixed portion 34b of the pedestal 34, that is, around the vertical axis parallel to the vertical axis of the fifth joint 39. It is possible. As a result, the robot arm 5 constitutes the multi-link manipulator having 6 degrees of freedom so as to be rotatable around a total of six axes.

Each joint that constitutes a rotating portion of each axis includes a motor 43 as an example of a rotation driving device and an encoder 44 that detects a rotation phase angle (that is, a joint angle) of the rotation axis of the motor 43. The motor 43 is provided in one member of a pair of members (for example, a rotation-side member and a support-side member that supports the rotation-side member) constituting each joint portion, and is described later. (Actually, it is arranged inside one member of each joint portion of the robot arm 5). The encoder 44 is provided in one member in order to detect the rotation phase angle (that is, the joint angle) of the rotation shaft of the motor 43 (actually, one member of each joint portion of the robot arm 5 is provided). Arranged inside). The rotating shaft of the motor 43 provided in one member is connected to the other member, and the other member can be rotated around each axis with respect to the one member by rotating the rotating shaft forward and backward. .

Reference numeral 46 denotes a rail coordinate system O _d , which indicates a relative positional relationship from the point O _s at the end of the rail 8 (see FIG. 8). Reference numeral 41 denotes a pedestal coordinate system of the pedestal 34 fixed to the rail movable portion 8b of the rail 8, and shows a relative positional relationship from the rail coordinate system _Od . The hand coordinate system 42 indicates a relative positional relationship from the platform coordinate system 41.

The origin position O _d (x, y) of the rail coordinate system 46 viewed from the end point O _s of the rail 8 is set as the position (rail position) of the rail movable portion 8b. Further, the origin position O _e (x, y, z) of the hand coordinate system 42 viewed from the base coordinate system 41 is set as the hand position of the robot arm 5 (the position of the tip of the suction nozzle 9 and the mop 10), and the base coordinates The posture of the hand coordinate system 42 viewed from the system 41 is expressed by the roll angle, the pitch angle, and the yaw angle (φ, θ, ψ) as the hand posture of the robot arm 5, and the hand position and posture vector are expressed as a vector r = [ x, y, z, φ, θ, ψ] ^T. The roll angle, pitch angle, and yaw angle will be described with reference to FIGS. 13A to 13C.

First, consider a coordinate system in which the coordinate system is rotated by an angle φ with the Z axis of the absolute coordinate system 35 as the rotation axis (FIG. 13A). The coordinate axes at this time are assumed to be [X ′, Y ′, Z].

Next, this coordinate system is rotated around the Z axis by an angle θ with Y ′ as the rotation axis (see FIG. 13B). The coordinate axes at this time are [X ″, Y ′, Z ″].

Finally, this coordinate system is rotated around the X ″ axis by an angle ψ using the X ″ axis as a rotation axis (see FIG. 13C). The coordinate axes at this time are [X ″, Y ′ ″, Z ′ ″]. The posture of the coordinate system at this time is a roll angle φ, a pitch angle θ, and a yaw angle ψ, and the posture vector at this time is (φ, θ, ψ). When the coordinate system of the posture (φ, θ, ψ) matches the coordinate system obtained by translating the origin position to the origin position O _e (x, y, z) of the hand coordinate system 42, and the hand coordinate system 42, Assume that the orientation vector of the coordinate system 42 is (φ, θ, ψ).

When controlling the hand position and orientation of the robot arm 5 respectively, the hand position and orientation vector r, to be made to follow on the hand position generated by the target track generation unit 55 to be described later and orientation target vector r _d Become.

Reference numeral 26 denotes a data input IF (interface). A person (housework worker) uses a button or a keyboard or an input device such as a mouse or a microphone to input a command to start or end the cleaning work to the housework robot 1. Interface.

The display unit 14 is, for example, a display device installed on the work table 7, and displays on the display unit 14 the type of a robot operation or a parameter to be corrected, which will be described later.

The operation database 17 stores and holds information (information related to the cleaning operation) regarding the movement of the rail movable unit 8b and the robot arm 5 such as the position and posture at a certain time. Here, the information related to the cleaning operation includes information on the position and orientation of the hand of the robot arm 5 according to the cleaning operation performed by the robot arm 5, information on the force applied by the robot arm 5 to the device 6, and information on the robot arm 5 6 includes at least one information of information on the strength of the suction force when sucking 6, speed information of the robot arm 5, and work impossible area information which is information on an area where cleaning is not performed.

Details of the operation database 17 will be described.

The operation database 17 is, for example, information on the operation of the rail movable unit 8b and the robot arm 5 shown in FIG. 4, and a work ID number for identifying a cleaning work, and an operation ID number for identifying individual actions in the work. , Information on the position of the rail movable part 8b in the operation, information on the hand position and posture of the robot arm 5 in the operation, information on the force applied to the work surface by the robot arm 5 in the operation, and strong suction force Information on the height, information on a flag (validity flag) indicating whether or not any of the parameters of the position, posture, force, and suction force of the robot arm 5 is valid, and each action acts Information on time and the type of parameter to be corrected when correcting the operation information in the operation database 17 by the operation correction unit 20 described later And information about, configured to hold and progress information indicating whether the currently operating.

The work ID number for identifying the cleaning work in the operation database 17 is information indicating the work ID number assigned to each cleaning work in order to identify each other when there are a plurality of types of cleaning work.

The operation ID number for identifying each operation in the cleaning work in the operation database 17 identifies each cleaning operation in one cleaning operation from each other when one cleaning operation is composed of a plurality of cleaning operations. Therefore, it is information representing the operation ID number assigned to each cleaning operation.

The information on the position of the rail movable portion 8b in the operation database 17 represents the information on the rail position described above, that is, the origin position O _d (x, y) of the rail coordinate system 46 viewed from the O _s at the end of the rail 8. For example, as shown in FIG. 8, when the housework robot 1 performs a cleaning operation by running from left to right on the rail fixing portion 8 a, the first rail position (x ₁ , y ₁₎ of the rail movable portion 8 b is performed. ), The second rail position (x ₂ , y ₂ ), the third rail position (x ₃ , y ₃ ), and the like.

Information on the position of the rail movable portion 8b in the operation database 17 is set in advance in the operation database 17, or the robot arm 5, the suction nozzle 9 or the mop 10 is directly held by the human hand 16 and will be described later. The robot arm 5 may be moved and stored in the impedance control mode.

The information on the hand position and posture of the robot arm 5 in the motion database 17 represents the hand position and posture of the robot arm 5 described above, and from the origin position O _e and the posture, (x, y, z, φ, θ , Ψ).

For example, as shown in FIG. 9, the robot database 5 or the suction nozzle 9 or the mop 10 is directly gripped by the human hand 16 and the impedance described later. In the control mode, the robot arm 5 is moved, and information on the hand position and posture of the robot arm 5 (the dotted line path in FIG. 9) is acquired by the control unit 22 at certain time intervals (for example, every 0.2 msec). (Specifically, as described in the explanation of the control unit 22, the joint angle measured by the encoder 44 of each joint unit by the forward kinematics calculation unit 58 is converted into the hand position and posture, and the robot arm 5 Information on the hand position and posture is acquired) and stored together with the time information in the motion database 17 in the motion database 17. It should be noted that information on position, posture and time may be generated in advance by the manufacturer at the time of product shipment and stored in the operation database 17. Further, the robot arm 5 is moved, and the environment (environment including the robot arm 56 and the device 6) is photographed with an image pickup device such as a camera (for example, disposed above the robot arm 5), and the obtained image A model matching process is performed between data (for example, the image of the device 6 in the obtained environment information) and an object image (for example, an image of the device 6) stored in advance, and the matching position is determined by the robot arm 5. Although it is not specifically shown, it may be stored in the operation storage unit 15 through the data input IF 26 as the hand position of the user.

The information on the force applied by the robot arm 5 stored in the motion database 17 indicates information on the force applied to the target object when the robot arm 5 performs work, and the robot arm 5 has x, y, and z directions. Forces to be applied are f _x , f _y , and f _z , respectively, and forces to be applied in the φ, θ, and ψ directions are f _φ , f _θ , and f _ψ . In operation database _17, expressed as _{_{(f x, f y, f}} z, f φ, f θ, f ψ). For example, when f _z = 5 [N], this means that the work is performed with a force of 5N in the z-axis direction, and when the device 6 is wiped and cleaned, the device 6 is rubbed with a force in the vertical direction. It is a parameter used for.

The information on the suction force in the operation database 17 indicates the suction force when the robot arm 5 performs the suction work. The suction forces in the x, y, and z directions of the robot arm 5 are p _x , p _y , and p _z , respectively, and the suction forces in the φ, θ, and ψ directions are p _φ , p _θ , and p _ψ . In the operation database 17, it is expressed as (p _x , p _y , p _z , p _φ , p _θ , p _ψ ). For example, as the value of p increases, the suction force increases. For example, when a large amount of dust is attached, the suction force is set to a large value (for example, set to a value of “5”). The force is set small (for example, set to a value of “2”).

Information on a flag (validity flag) indicating whether or not any of the parameters of the position, posture, force, and suction force of the robot arm 5 in the motion database 17 is valid, that is, the motion of FIG. The flag information in the database 17 is a value indicating which information among the position, posture, force, and suction force of the robot arm 5 indicated by each action ID is valid. Specifically, the information is shown in FIG. It is expressed as a 32-bit numerical value. In FIG. 5, “1” is set when the values of the position, posture, force, and suction force are valid for each bit, and “0” is set when the values are invalid. For example, the 0th bit is “1” if the x-coordinate value of the position is valid, and “0” if it is invalid. The first bit is “1” if the y-coordinate value of the position is valid, and “0” if it is invalid. The second bit is “1” when the position z-coordinate value is valid, and “0” when the position is invalid. The third, fourth, and fifth bits sequentially indicate the validity of the posture φ, θ, and ψ. To express. The sixth to eleventh bits indicate whether the force components f _x , f _y , f _z , f _φ , f _θ , and f _ψ are valid or invalid. The 12th to 17th bits indicate whether the components p _x , p _y , p _z , p _φ , p _θ , and p _ψ of the attractive force are valid or invalid. In addition, since many flags (32 bits) are prepared for future expansion in this example, since the 18th to 31st bits are not used, “0” is inserted, but only the 18th bit is used. Or a variable that can be stored. In Figure 5, 0 1 bit from bit is "1", since the 8th bit is "1", among the operations information, the position of x, only the f _z information y information and force are valid It is assumed that any value stored in the values other than z, φ, θ, ψ and force f _z and the suction force in the operation information is invalid.

Information on the time at which each action in the action database 17 acts, that is, the time in the action database 17 in FIG. 4 is the time required for the housework robot 1 to execute each action, and the action stored in this action ID. Represents that the housework robot 1 operates over the time stored here. This time represents the relative time from the previous operation, not the absolute time. That is, the time until the rail movable part 8b and the robot arm 5 move is represented by the position of the rail movable part 8b and the position and posture of the robot arm 5 indicated by the operation ID.

Information relating to the type of parameter to be corrected when the motion correction unit 20 corrects the motion information in the motion database 17 in the motion database 17, that is, the correction parameter flag in FIG. This is information indicating which parameter is to be corrected according to the type. Specifically, it is represented by a 32-bit numerical value shown in FIG. In FIG. 6, “1” is set when each value of position, posture, force, and suction force can be corrected by each bit, and “0” is set when correction is impossible. For example, the 0th bit is “1” if the x-coordinate value of the position can be corrected, and “0” otherwise. The first bit is “1” if the y-coordinate value of the position can be corrected, and “0” otherwise. The second bit is “1” if the z-coordinate value of the position can be corrected, and “0” otherwise. The third, fourth, and fifth bits sequentially indicate the possibility of correcting the postures φ, θ, and ψ. Similarly, the 6th to 11th bits indicate the correctability of the force, and the 12th to 17th bits indicate the correctability of each component of the suction force. In addition, since many flags (32 bits) are prepared for future expansion in this example, since the 18th to 31st bits are not used, “0” is inserted, but only the 18th bit is used. Or a variable that can be stored.

The progress information indicating whether or not the operation database 17 is currently operating is information indicating whether or not the housework robot 1 is currently operating. When the operating database 17 is operating, “1” is recorded and operating. Otherwise, “0” is recorded. Specifically, a person selects a cleaning operation to be performed via the data input IF 26, and the selected information is input from the data input IF 26 to the operation selection unit 29. When the first cleaning operation of the selected work is started by the housework robot 1, the operation selection unit 29 sets “1” for the operation currently in operation among the plurality of operations constituting the work. 17, and “0” is stored in the operation database 17 for operations that are not operating. The information indicating whether or not the operation is in progress is input to the operation storage unit 15 through the operation correction unit 20 as a notification that the operation instructed from the control unit 22 has ended, and the operation storage unit 15 stores the information in the operation database 17. Remember. In the operation database 17, regarding the progress information, whether or not the operation has ended is determined by measuring the time with the built-in timer of the control device 1000 and determining the end of the operation.

3 selects the optimum cleaning work from the work list of the action database 17 (for example, work indications such as “wiping” and “suction” displayed in the center of the cleaning switch 26c in FIG. 26). When 16A selects via the data input IF 26, “1” is set in the progress information of the currently operating operation ID among the selected operations and stored in the operation database 17; “0” is set and stored in the operation database 17.

The unworkable area database 28 stores information related to areas where the housework robot 1 does not perform work (wipe cleaning work or suction work). Specific information is shown in FIG. In FIG. 10, the position (x, y) of the work disabling region represents a region where a person does not want the housework robot 1 to perform a wiping operation or a suctioning operation. For example, if the hatched area of the cleanable surface R in FIG. 11 is the work non-operation area RB, the coordinates necessary to represent the area RB (in this example, the coordinates of the four corners of the rectangular area ( x _c1 , y _c1 ), (x _c2 , y _c2 ), (x _c3 , y _c3 ), (x _c4 , y _c4 )) are stored. Note that each of the coordinates, of the working path of the cleaning area RA to perform cleaning, represented by relative coordinates from the coordinates O _s of the rail. Coordinates representing these unworkable areas RB are generated by an operation correction unit 20 described later and stored in the unworkable area database 28.

The correction operation type determination unit 23 determines a type of correction that can be corrected by a person applying force to the robot arm 5 with his / her hand 16 in the operation correction unit 20 described later. For example, as shown in FIG. 12C, when a person applies force to the robot arm 5 from the lateral direction with his hand 16, the direction parallel to the work surface of the robot arm 5 (for example, when the work surface is along the horizontal direction, In the following description, in order to simplify the description, the cleaning area RA can be translated by simply moving the position of “horizontal direction”. The type of correction operation in this case is “movement of the position of the work surface”. When a person applies a downward force to the robot arm 5 from above the robot arm 5 with his / her hand 16 as shown in FIG. 27B while wiping and cleaning the device 6 with the robot arm 5 as shown in FIG. The correction unit 20 can set the force applied during cleaning as shown in FIG. 27C. The type of the correction operation in this case is “force application”. As described above, the correction operation type determination unit 23 can determine the type of correction of the cleaning operation from the degree of force applied to the robot arm 5 by a human hand and the hand position of the robot arm 5. Details will be described later.

The motion correction unit 20 cleans the motion database 17 by applying a force to the robot arm 5 with the hand 16 while the housework robot 1 is performing a cleaning operation based on the position, posture, and time information of the motion database 17. It has a function to correct operation information. Details will be described later.

The operation storage unit 15 stores the operation information corrected by the operation correction unit 20 in the operation database 17 or the work impossible area database 28.

Next, details of the control parameter management unit 21 will be described.

Based on the motion correction instruction from the motion correction unit 20, the control parameter management unit 21 performs impedance control mode, hybrid impedance control mode, force control mode, force hybrid impedance control mode, and high-rigidity position of the robot arm 5. Setting of switching between control modes, setting of mechanical impedance setting values in each control mode, and hand position and posture target correction output _rdΔ output by the impedance calculation unit 51 of the control unit 22 in each control mode Setting and setting of operation information to the target trajectory setting unit 55 of the control unit 22 are performed.

Further, the control parameter management unit 21 detects the position of the rail movable unit 8b stored in the cleaning information database 17 (the origin position O _d (x, y) of the rail coordinate system 46 as viewed from the coordinates O _s of the end of the rail 8). ) To generate a route in the cleaning area RA excluding the work impossible area RB in the work impossible area database 28. Also, the control parameter management unit 21 receives information such as the hand position or force information of the robot arm 5 from the control unit 22, and notifies the operation correction unit 20 of the information. When a command for opening / closing the hand 30 is input by the data input IF 26, input information from the data input IF 26 is input to the hand control unit 54 of the control unit 22 via the control parameter management unit 21, and control parameter management is performed. An opening / closing command for the hand 30 is issued from the unit 21 to the hand control unit 54.

The position control mode is a mode in which the robot arm 5 operates based on a hand position and posture target vector command of a target trajectory generation unit 55 described later.

The impedance control mode is a mode in which the robot arm 5 operates according to the force applied to the robot arm 5 from a person or the like.

The hybrid impedance control mode is a mode (impedance control mode) in which the robot arm 5 operates according to the force applied to the robot arm 5 from a person or the like while the robot arm 5 is operating in the position control mode. In this mode, the position control mode and the impedance control mode are performed simultaneously. For example, during a cleaning operation for sucking dust or the like on the work surface, as shown in FIG. 12B, the robot arm 5 is directly held by a person's hand 16 and correction is performed such that the cleaning area RA is translated.

The force control mode is a control mode in which the robot arm 5 operates while pressing the suction nozzle 9 or the mop 10 against the work surface with a force applied in advance to the control unit 22. For example, the robot arm 5 moves against the work surface. This is a control mode used for a work surface component of the robot arm 5 when performing a cleaning operation to wipe off dirt by applying a certain force.

The force hybrid impedance control mode is a control mode that switches between the hybrid impedance control mode and the impedance control mode for each of the six axes, and further operates in a force control mode that operates by applying a specified force. Note that the impedance control mode cannot be set in the direction in which the force control mode is set (the force control mode and the impedance control mode are in an exclusive relationship).

These control modes are operated by setting appropriate control modes as follows for each direction and posture of the robot arm 5 during operation.

For example, when the housekeeping robot 1 performs a wiping operation by applying a specified force vertically downward to the work surface while moving in a circle parallel to the work surface of the device 6 as shown in FIG. Set the control mode. Specifically, the following control modes are set for each of the six axes (x, y, z, φ, θ, ψ). That is, the (x, y) component operates in the hybrid impedance control mode, the (φ, θ, ψ) component operates in the impedance control mode, and the z-axis component operates in the force control mode. . In this way, the direction parallel to the work surface of the device 6 is set to the hybrid impedance control mode, so that it is switched to the hybrid impedance control mode when a person is operating while operating in the position control mode. Thus, the robot arm 5 can be moved according to the force applied to the robot arm 5 from a person or the like. Further, by setting the (φ, θ, ψ) components in the impedance control mode, the posture of the robot arm 5 can be changed according to the force applied to the robot arm 5 from a person or the like while stopped. become. Further, by setting the z-axis component to the force control mode, it becomes possible to operate while pressing with a designated force.

Similarly, the force hybrid impedance control mode is set even when the housework robot 1 moves in a circular shape parallel to the work surface as shown in FIG. 23 and sucks and cleans the dust on the work surface. Specifically, the (x, y) component is operated in the hybrid impedance control mode, the (φ, θ, ψ) component is operated in the impedance control mode, and the z-axis component is operated in the force control mode.

The high-rigidity position control mode is a mode in which the position control mode during the cleaning operation is further increased in rigidity, and is realized by increasing the gain in the position error compensation unit 56 described later. When the robot arm 5 is applied to the robot arm 5, the force applied by the human hand 16 can be detected by the force detection unit 53 based on the amount of change in the hand position of the robot arm 5 by making the robot arm 5 unable to move easily. it can.

設定 Inertia M, viscosity D, and rigidity K are set as mechanical impedance setting parameters. Each parameter of the mechanical impedance set value is set based on the following evaluation formula using the correction value.

In the equations (1) to (3), KM, KD, and KK are gains, each of which is a constant value.

The control parameter management unit 21 outputs the inertia M, viscosity D, and rigidity K of the mechanical impedance parameters calculated based on the equations (1) to (3) to the control unit 22, respectively.

According to the above equations (1) to (3), for example, as shown in FIG. 12C, when a person wants to make a correction so as to move the area of the work surface, position components and posture components other than the x-axis and y-axis can be simplified. If it moves to the point, it becomes difficult to perform the correction operation. Therefore, the viscosity D and stiffness are set by setting the above correction value high (specifically, about 10 times higher) in the control parameter management unit 21 only for position components and posture components other than the x-axis and y-axis. Since K is set to be large, resistance or hardness is generated in the movement of the robot arm 5, and it is difficult for the position component and posture component other than the x-axis and the y-axis to move.

Alternatively, as another method, among the components of the hand position and posture target correction output _rdΔ output from the impedance calculation unit 51 described later, all values other than the x-axis and the y-axis are obtained by the control parameter management unit 21. There is a way to make it zero. As a result, the parts other than the x-axis and the y-axis cannot be moved by the force of the human hand 16, so that an erroneous operation can be prevented.

Further, it is necessary to notify the motion correction unit 20 of the hand position and posture of the robot arm 5 and information on the force applied by the person (information on the force of the person acting on the robot arm 5) from the control parameter management unit 21. is there. Therefore, the control parameter management unit 21 receives information on the hand position and force of the robot arm 5 from the control unit 22, and moves to the operation selection unit 29, the operation storage unit 15, and the operation correction unit 20 from the control parameter management unit 21. Make a notification. In addition, the control parameter management unit 21 notifies the control unit 22 of operation information such as position, posture, and time input from the operation correction unit 20.

FIG. 7 shows a block diagram of the control unit 22. The control unit 22 operates in the control mode set by the control parameter management unit 21, and further, according to the control mode, the mechanical impedance of the robot arm 5 set based on the set values of the inertia M, the viscosity D, and the stiffness K. The mechanical impedance value of the robot arm 5 is controlled to the set value. Furthermore, in the case of cleaning that performs suction, the control unit 22 performs control to rotate the rotating brush 11 while suctioning with a specified suction force. Furthermore, the control part 22 performs control which presses a work surface with the designated force in the case of wiping and cleaning. Furthermore, the control part 22 controls the rail movable part 8b, and performs control which moves the robot arm 5 to the designated position on the rail fixing | fixed part 8a.

Next, details of the control unit 22 will be described with reference to FIG.

The control unit 22 includes a robot arm control unit 49 that controls driving of the motor 43 of each joint unit of the robot arm 5, a rotating brush control unit 13 that controls driving of the motor 69 of the rotating brush 11, and a rail movable unit 8b. And a rail control unit 48 for controlling the driving of the motor 65. The robot arm control unit 49 includes a position error calculation unit 50, an impedance calculation unit 51, a force detection unit 53 as an example of a force detection unit, a hand control unit 54, a target trajectory generation unit 55, a position error compensation unit 56, and an approximate reverse motion. A scientific calculation unit 57 and a forward kinematics calculation unit 58 are provided. The position error compensation unit 56, approximate inverse kinematics calculation unit 57, and forward kinematics calculation unit 58 constitute a position control system 59.

Next, the robot arm control unit 49 will be described in detail.

From the robot arm 5, the current value (joint angle vector) vector q = [q ₁ , q ₂ , q ₃ , q ₄ , q ₅ , q of the joint angle measured by the encoder 44 of the joint axis of each joint part. ₆ ] ^T is output and taken into the control unit 22 by the input / output IF 24. However, q ₁ , q ₂ , q ₃ , q ₄ , q ₅ , q ₆ are respectively the first joint part 35, the second joint part 36, the third joint part 37, the fourth joint part 38, and the fifth joint. This is the joint angle of the part 39 and the sixth joint part 40.

Target track generation unit 55 receives an input of the cleaning operation from the control parameter managing unit 21, for implementing the operation of the robot arm 5 to the target, and the hand position and orientation target vector r _d, the hand force vector _fd and a flag indicating which parameter is effective for each direction (a flag indicating validity) are output. The operation of the target robot arm 5 is performed from the operation correction unit 20 via the control parameter management unit 21 according to the target cleaning operation, for each time (t = 0, t = t ₁ , t = t ₂ ,...) And position information and posture information (r _d0 , r _d1 , r _d2 ,...) And force information (f _d0 , f _d1 , f _d2 ,...) , Information of suction force (p _d0 , p _d1 , p _d2 ,...) Is given to the target trajectory generation unit 55.

Target track generation unit 55, generates using a polynomial interpolation, and the track between points, and the force, by interpolating a suction force, the tip unit position and orientation target vector r _d and the force vector f _d and the suction force p _d To do.

The hand control unit 54 instructs the hand opening / closing driving motor 62 of the robot arm 5 to open / close the hand 30 by driving the hand opening / closing driving motor 62 by the hand opening / closing command input from the control parameter management unit 21. Put out.

The force detection unit 53 functions as an example of a force detection unit, and detects an external force applied to the robot arm 5 due to contact between a person or the like and the robot arm 5. The force detection unit 53 includes a current value i = [i ₁ , i ₂ , i ₃ , i ₄ , i, which flows through the motor 43 that drives each joint unit of the robot arm 5 and is measured by a current sensor of the motor driver 47. ₅ , i ₆ ] ^T is taken in via the input / output IF 24, and the current value q of each joint angle of each joint is taken in via the input / output IF 24, and from an approximate inverse kinematics calculation unit 57 described later. The joint angle error compensation output u _qe is _fetched . The force detection unit 53 functions as an observer, and the torque generated in each joint by external force applied to the robot arm 5 based on the current value i, the current value q of the joint angle, and the joint angle error compensation output u _qe. τ _ext is calculated. _{_{^{Then, F ext = J v (q}}} ) -T τ ext - [0,0, mg] in terms of the equivalent tip unit external force _{F ext} at the tip unit of the robot arm 5 and outputs the equivalent tip unit external force _{F ext} by ^T. Where J _v (q) is

Jacobian matrix that satisfies However, v = [v _x , v _y , v _z , ω _x , ω _y , ω _z ] ^T , and (v _x , v _y , v _z ) represents the robot arm 5 in the hand coordinate system 42. The translation speed of the hand, (ω _x , ω _y , ω _z ) is the angular velocity of the hand of the robot arm 5 in the hand coordinate system 42. Further, m is the weight of the

cleaning unit

9 or 10 attached to the hand 30 of the robot arm 5, and g is the gravitational acceleration. The value of the weight m of the

cleaning unit

9 or 10 can be input by the person from the data input IF 26 to the force detection unit 53 before attaching the

cleaning unit

9 or 10, but normally the cleaning unit suction nozzle 9 or mop Since the weight m of 10 is not a value that is frequently changed, it may be a preset value.

The impedance calculation unit 51 is a part that performs the function of realizing the control of the mechanical impedance value of the robot arm 5 to the mechanical impedance set value.

When the impedance control mode is designated, the hand position and posture target correction output _rdΔ is output from the impedance calculator 51. When switching to the force hybrid impedance control mode, if there is a force component designated as valid by a flag (a flag indicating validity), inertia that is an impedance parameter set by the control parameter management unit 21 Based on M, viscosity D, rigidity K, current value q of the joint angle, external force F _ext detected by the force detector 53, and f _d output from the target trajectory generator 55, the robot arm 5 A hand position and posture target correction output _rdΔ for realizing control that causes the machine impedance value to approach the machine impedance set value for the robot arm 5 is calculated by the impedance calculation unit 51 using the following equation (4). Output from the impedance calculator 51.

Hand position and orientation target correcting output r _d? Is added by the position error calculation unit 50 to the hand position and orientation target vector rd outputted by the target track generation unit 55, the tip unit position and orientation correction target vector r _dm is the position error calculation unit 50 is generated. For example, in order to perform cleaning by applying force only in the z-axis direction and move the other components in the position control mode, the position error calculation unit 50 sets the hand position and posture target correction output _rdΔ other than the z component to 0. Set with.

However,

And s is a Laplace operator.

Position error calculation unit 50 further calculates the tip unit position and orientation correction target vector r _dm, error r _e between the tip unit position and orientation vectors r calculated by the forward kinematics calculation unit 58 to be described later, the determined error r _e is output to the position error compensator 56.

The forward kinematics calculator 58 receives a joint angle vector q, which is a current value q of the joint angle measured by the encoder 44 from the encoder 44 of each joint axis of each joint of the robot arm 5 via the input / output IF 24. Entered. The forward kinematics calculator 58 performs a geometric calculation of conversion from the joint angle vector q of the robot arm 5 to the hand position and posture vector r. The hand position and posture vector r calculated by the forward kinematics calculator 58 is output to the position error calculator 50, the impedance calculator 51, and the target trajectory generator 55.

Positional error compensating unit 56, based on the error r _e obtained by the position error calculation unit 50, and outputs a positional error compensating output u _re the approximation reverse kinematical calculation unit 57.

Specifically, the position error compensation output _ure is

Is calculated by Here, _{K P} is a proportional gain matrix, _{K I} is an integral gain matrix, _{K d} is the derivative gain matrix, diagonal components hand position vector _{r e = [x, y,} z, φ, θ, ψ] It is a diagonal matrix composed of gains for each component of ^T.

Further, when the high-rigidity position control mode is set, the position error compensation unit 56 sets the proportional gain matrix K _P , the integral gain matrix K _I , and the differential gain matrix K _D to large values set in advance. Here, the high rigidity means that the rigidity is higher than that in the normal position control, and specifically, a large value is set as compared with the normal position control mode. For example, if the value is set to about twice that in the normal position control mode, the rigidity can be increased to about twice.

In this way, highly rigid position control can be realized. Note that by changing the value of the gain for each component, for example, it is possible to control so that only the z-axis direction is highly rigid and the other directions are operated by normal position control.

The approximation reverse kinematical calculation unit 57, based on the joint angle vector q that is measured at the positional error compensating output u _re and the robot arm 5 inputted from the positional error compensating unit 56, the approximation equation u _{out = J} r _(q ) Approximate inverse kinematics with ^-1 u _in . However, J _r (q) is

The Jacobian matrix satisfying the above relationship, u _in is an input to the approximate inverse kinematics calculation unit 57, u _out is an output from the approximate inverse kinematics calculation unit 57, and the input u _in is the joint angle error q _e Then, a conversion formula from the hand position / posture error r _e to the joint angle error q _e is obtained as q _e = J _r (q) ⁻¹ r _e .

Therefore, when the positional error compensating output u _re from the positional error compensating unit 56 is inputted to the approximation reverse kinematical calculation unit 57, as an output from the approximation reverse kinematical calculation unit 57, to compensate for the joint angle error q _e The joint angle error compensation output u _qe is output from the approximate inverse kinematics calculation unit 57 to the motor driver 25 of the robot arm 5 via the input / output IF 24.

The joint angle error compensation output u _qe is given as a voltage command value to the motor driver 25 of the robot arm 5 via the D / A board of the input / output IF 24, and each joint axis is driven to rotate forward and reverse by each motor 43. 5 operates.

The rotating brush control unit 13 controls to rotate the rotating brush 11 by driving and controlling the motor 69 of the rotating brush 11 according to the suction force input from the target trajectory generating unit 55.

The rail control unit 48 drives and controls the motor 65 of the rail movable unit 8b based on the positional information of the rail movable unit 8b input from the target track generation unit 55, and moves the robot arm 5 on the rail fixed unit 8a. Control is performed so as to move together with 8b. Specifically, the rail controller 48 controls forward / reverse rotation driving of the motor 65 of the rail movable portion 8b, and the rail movable portion 8b to which the robot arm 5 is connected can be moved in the left-right direction on the rail fixing portion 8a. It is said.

Hereinafter, actual operation steps of the robot arm control program of the robot arm 5 will be described based on the flowchart of FIG.

The joint angle data (joint variable vector or joint angle vector q) measured by the encoders 44 of the joints of the robot arm 5 is taken into the control device body 45 (step S51).

Next, the inverse kinematics calculation unit 57 calculates the Jacobian matrix _{Jr and the} like necessary for the kinematics calculation of the robot arm 5 (step S52).

Next, the forward kinematics calculator 58 calculates the current hand position and posture vector r of the robot arm 5 from the joint angle data (joint angle vector q) from the robot arm 5 (step S53).

Then, based on the operation information transmitted from the operation correction unit 20, the target track calculation unit 55 calculates the tip unit position and orientation target vector r _d and the force target vector f _d of the robot arm 5 (step S54).

Next, the force detection unit 53 calculates an equivalent hand external force F _ext at the hand of the robot arm 5 from the drive current value i of the motor 43, the joint angle data (joint angle vector q), and the joint angle error compensation output u _qe. Calculate (step S55).

Next, in step S56, the control mode set by the control parameter management unit 21 is set. In the case of only the high rigidity position control mode, the process proceeds to step S57. On the other hand, in the case of the force hybrid impedance control mode, the impedance control mode, or the hybrid impedance control mode, the process proceeds to step S58.

In step S57 (processing in the impedance calculation unit 51), when the high rigidity position control mode is set in the control parameter management unit 21, the impedance calculation unit 51 sets the hand position and posture target correction output _rdΔ to 0. Let it be a vector. Thereafter, the process proceeds to step S59.

When the force hybrid impedance control mode, the impedance control mode, or the hybrid impedance control mode is set in the control parameter management unit 21, the inertia M, viscosity D, and rigidity of the mechanical impedance parameter set in the control parameter management unit 21 are set. From the K, joint angle data (joint angle vector q), and the equivalent hand external force F _ext applied to the robot arm 5 calculated by the force detection unit 53, the hand position and posture target correction output _rdΔ is obtained as an impedance calculation unit 51. (Step S58).

Then, the position in the error calculating unit 50, and the tip unit position and orientation correction target vector r _dm is the sum of the tip unit position and orientation target vector r _d and the tip unit position and orientation target correcting output r _d?, The current hand position and orientation vector r error _{r e} difference tip unit position and orientation is a is calculated (step S59, the step S60). In step S60, a PID compensator can be considered as a specific example of the position error compensator 56. Control is performed so that the position error converges to 0 by appropriately adjusting three gains of proportionality, differentiation, and integration, which are constant diagonal matrices. In step S59, high gain position control is realized by increasing the gain to a certain value.

In step S59 or step S61 subsequent to step S60, in the approximation reverse kinematical calculation unit 57, by multiplying the inverse matrix of the Jacobian matrix J _r calculated in step S52 in the approximation reverse kinematical calculation unit 57, the position error compensation output The approximate inverse kinematics calculation unit 57 converts u _re into a joint angle error compensation output u _qe that is a value related to a joint angle error from a value related to a hand position and posture error.

Following step S61, the joint angle error compensation output u _qe is given from the approximate inverse kinematics calculation unit 57 to the motor driver 25 via the input / output IF 24, and the amount of current flowing through each motor 43 is changed to change the robot arm 5. Rotational motions of the respective joint axes are generated (step S62).

The above steps S51 to S62 are repeatedly executed as a control calculation loop, thereby controlling the operation of the robot arm 5, that is, controlling the mechanical impedance value of the robot arm 5 to the appropriately set value. Can be realized.

Next, the correction operation type determination unit 23 and the operation correction unit 20 will be described in detail.

The correction operation type determination unit 23 determines the type of correction in which the operation correction unit 20 can correct the housework operation by applying force to the robot arm 5 with the human hand 16. There are the following seven types of correction.

The first correction type is “movement of the position of the work surface”. Specifically, as shown in FIG. 12A or 12B (viewed from above in FIG. 12A), while the device 6 is being cleaned in the position control mode with the robot arm 5, the robot with the human hand 16 as shown in FIG. 12C. When a force is applied to the arm 5 from the lateral direction, the cleaning region RA can be translated by moving the horizontal position of the robot arm 5 with respect to the work surface as shown in FIG.

The second type of correction is “force applied” to the work surface of the device 6 at the time of wiping and cleaning. This is effective when the force bit is “1” in the operation flag (the flag indicating validity) of the operation currently being performed (the progress information of the operation database 17 is “1”). As shown in FIG. 27A, while the work surface of the device 6 is being wiped and cleaned with the robot arm 5, as shown in FIG. As shown in FIG. 27C, when the force applied during wiping and cleaning is increased, and conversely, when the robot arm 5 is applied upward from below with the human hand 16, the force applied during wiping and cleaning is corrected to be weaker. be able to.

The third type of correction is “suction force” on the work surface of the device 6 during suction cleaning. This is effective when the bit of the suction force is “1” in the operation flag (the flag indicating validity) of the operation currently being performed (the progress information in the operation database 17 is “1”). As shown in FIG. 28A, when the work surface of the device 6 is suction-cleaned by the robot arm 5 and a force is applied downward from above to the robot arm 5 with a human hand 16 as shown in FIG. Thus, as shown in FIG. 28C, the suction force at the time of suction cleaning is set to be strong, and conversely, if the human hand 16 applies an upward force from below to the robot arm 5, the suction force at the time of suction cleaning is set to be weak. be able to.

The fourth type of correction is the movement “speed” of the hand of the robot arm 5 (ie, the cleaning units 9 and 10). While cleaning the work surface of the device 6 with the robot arm 5 as shown in FIG. 29A or 29B (viewed from the upper side of FIG. 29A), the person moves in the direction opposite to the moving direction of the robot arm 5 as shown in FIG. 29C. When the force is applied to the robot arm 5 with the hand 16, the operation correction unit 20 can reduce the speed during cleaning as shown in FIG. 29D. On the contrary, when the human hand 16 applies a force to the robot arm 5 with the human hand 16 in accordance with the traveling direction of the robot arm 5 while cleaning the work surface of the device 6 with the robot arm 5, the motion correcting unit 20 , Can speed up cleaning.

The fifth type of correction is “change in direction (posture)”. As shown in FIG. 30A or FIG. 30B (viewed from above in FIG. 30A), while cleaning the work surface of the device 6 with the robot arm 5, the longitudinal direction of the

cleaning units

9, 10 is illustrated with the human hand 16. When a force is applied in the direction in which the robot arm 5 is to be moved in a zigzag manner as in 30C, the motion correction unit 20 can change the traveling direction of the robot arm 5 during cleaning as shown in FIG. 30D. . This can be realized by changing the posture (φ, θ, ψ) of the hand of the robot arm 5.

The sixth type of correction is “area that you do not want to work on”. As shown in FIG. 31 with the hand 16 of the person 16A, the robot arm 5 (or the cleaning units 9 and 10) is forced along the outline of the region RB where the robot arm 5 is not desired to be operated. When the robot arm 5 is moved over, the operation correction unit 20 can set an area RB that is not desired to be worked as shown in FIG.

The seventh type of correction is “movement in the vertical direction of the work surface”. While the work surface of the device 6 is being cleaned with the robot arm 5 as shown in FIG. 32A, the robot arm 5 is moved upward by applying a force upward to the robot arm 5 with a human hand 16 as shown in FIG. 32B. Then, for example, as illustrated in FIG. 32C, the operation correction unit 20 can clean the work surface 6 as of another device 6 a such as a cutting board disposed on the work surface of the device 6.

The correction operation type determination unit 23 determines one type of correction among the seven types of correction. Specifically, one of the seven types of correction is selected by the data input IF 26 such as a button, or is detected by the force detection unit 53 and acquired by the information acquisition unit 100. Information on the relationship between the force applied to the robot arm 5 by the human hand 16, the force applied to the robot arm 5 stored in the motion database 17 and acquired by the information acquisition unit 100, and the type of correction (for example, the direction in which the force is applied) And the relationship information between the magnitude and the correction type), the correction operation type determination unit 23 estimates the type.

Hereinafter, specific correction type estimation processing of the correction type estimation method will be described in detail with reference to the flowchart of FIG.

If the power button 26a of the household robot 1 is set to “ON” and the robot arm 5 is not gripped by the human hand 16 and no force is applied to the robot arm 5, the robot arm 5 does not move. When a force is applied to the robot arm 5 by the human hand 16, the robot arm 5 is moved in the direction in which the robot arm 5 is moved in the impedance control mode (a mode in which the force of the human hand 16 is detected and moved by impedance control). Can be made. In this case, the force detection unit 53 of the control unit 22 detects the force acting on the robot arm 5, and the information on the force detected by the force detection unit 53 is determined via the information acquisition unit 100. Input to the unit 23 (step S1).

Next, in step S2, all the components of the acquired in detected by the force detection unit 53 and the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is Whether or not it is below a certain threshold (specifically, (f _dx , f _dy , f _dz , f _dφ , f _dθ , f _dψ )) of ID “1” in FIG. Judge with. All components were detected by the force detection unit 53 and acquired by the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is below the certain threshold If it is determined by the correction operation type determination unit 23, the robot arm 5 does not move and is not corrected (step S20), and the correction type estimation process of the correction operation type estimation method ends. The control mode in that case is an impedance control mode.

In step S2, one of the components _(f _x of the acquired detected by the force detection unit 53 and the information acquisition unit 100 _{force, f y, f z, f} φ, f θ, of the six components of the f _[psi ) _{Exceeds a} certain threshold (specifically, (f _dx , f _dy , f _dz , f _dφ , f _dθ , f _dψ ) of ID “1” in FIG. 33). If the type determining unit 23 determines, the process proceeds to step S3.

In step S3, the corrected operation type determination unit 23 further determines whether the current housework robot 1 is operating in the operation database 17 based on the information acquired via the information acquisition unit 100. Specifically, the correction operation type determination unit 23 determines that no operation is selected by the operation selection unit 29 and the progress information is “0” for all the operation IDs in the operation database 17. In the case of determination (a state in which work is not started), the correction operation type determination unit 23 determines that the operation database 17 is not operating, and the process proceeds to step S6. When the operation selection unit 29 selects the cleaning operation and starts cleaning, and the correction operation type determination unit 23 determines that the progress information is “1”, the operation database 17 If it is operating, the correction operation type determination unit 23 determines that the operation is in progress, and the process proceeds to step S4.

In step S4, when the robot arm 5 is gripped by a human hand 16 and a force is applied in a direction in which the operation of the robot arm 5 is to be corrected, the force applied to the robot arm 5 is detected by the force detection unit 53, and the force is detected. Displacement of each of the forces (f _x , f _y , f _z , f _φ , f _θ , f _ψ ) detected by the detection unit 53 and acquired through the information acquisition unit 100 for a certain period of time. The amount is measured by the correction operation type determination unit 23, and it is determined whether the displacement amount of the position component (f _x , f _y , f _z ) or the posture component (f _φ , f _θ , f _ψ ) is larger. Measurement is performed by the determination unit 23. Specifically, as shown in FIG. 15, each time series force (f _x , f _y , f _z , f _φ , f _θ , f _ψ ) is measured by the correction operation type determination unit 23, and is constant. The correction operation type determination unit 23 measures how much the force is displaced in time (for example, time 1), and the correction operation type determination unit 23 measures the component having the largest displacement. In this example, the displacement of f _phi largest, as judged by the correction operation type determining unit 23 orientation component is force is applied from the position component, the process proceeds to step S9.

When the correction operation type determination unit 23 determines in step S4 that the displacement amount of the posture is larger than the displacement amount of the position, the correction operation type determination is made that the correction type is a “direction (posture) change” type. Then, the correction type estimation process is terminated (step S9). The control mode at that time is the same control mode (force hybrid impedance control mode) as before the correction type is determined.

On the other hand, when the correction operation type determination unit 23 determines in step S4 that the displacement amount of the position is equal to or greater than the displacement amount of the posture, the force component in the direction perpendicular to the work surface (for example, installed horizontally on the ground) In the case where the work surface of the device 6 is polished, f _z ) is determined by the correction operation type determination unit 23 to determine whether it is equal to or greater than a certain threshold (specifically, f _dz of ID “1” in FIG. 33). (Step S5).

In step S5, when it is determined by the correction operation type determination unit 23 that the force component in the direction perpendicular to the work surface is smaller than the certain threshold, the force component in the direction parallel to the work surface (for example, horizontal to the work table 7). _f x when cleaning the working surface of the device 6 set in, either or both of the _{f y)} is, the certain threshold (specifically, _f dx of ID in FIG. 33, _{"1", f dy)} or Is determined by the correction operation type determination unit 23 (step S10).

In step S10, _{(specifically,} f _x, f y of ID in FIG. 33, _"1") a certain threshold the horizontal direction force component to the working surface determined by the less than correcting operation type determination unit 23 If so, it is determined that there is no correction (no type), and the correction type estimation process is terminated (step S11). If there is no correction, stop the correction and work.

In step S10, when the correction operation type determination unit 23 determines that the force component in the direction horizontal to the work surface is equal to or greater than the certain threshold value, the process proceeds to step S13.

In step S13, the horizontal movement amount of the work surface calculated by the correction operation type determination unit 23 is greater than or equal to a certain threshold (specifically, g _x , g _y of ID “2” in FIG. 33). When the correction operation type determination unit 23 determines, the correction operation type determination unit 23 determines the type of “movement of the position of the work surface” as the correction type, and ends the correction type estimation process (step S14). . In addition, when the movement amount in the horizontal direction of the work surface is calculated by the correction operation type determination unit 23, specifically, the robot before a human operation from the control unit 22 via the control parameter management unit 21 or the information acquisition unit 100 is used. The hand position of the arm 5 and the hand position during operation are input to the correction operation type determination unit 23, and a value obtained by subtracting the hand position before operation from the hand position during operation is calculated by the correction operation type determination unit 23. be able to. In addition, when the movement amount in the direction perpendicular to the work surface is calculated by the correction operation type determination unit 23, specifically, from the control unit 22 via the control parameter management unit 21 or the information acquisition unit 100 before the human operation. A value obtained by inputting the z component of the hand position of the robot arm 5 and the z component of the hand position during operation to the correction operation type determination unit 23 and subtracting the z component of the hand position before operation from the z component of the hand position during operation. Can be calculated by the correction operation type determination unit 23 as a movement amount.

In step S13, when the correction operation type determination unit 23 determines that the horizontal movement amount of the work surface is less than the certain threshold value, the “speed” in the direction horizontal to the work surface is set as the correction type. The type is determined, and the correction type estimation process is terminated (step S15).

In step S5, when the correction operation type determination unit 23 determines that the force perpendicular to the work surface is equal to or greater than the certain threshold, the vertical direction of the work surface calculated by the correction operation type determination unit 23 is further determined. The correction operation type determination unit 23 determines whether or not the movement amount is larger than a certain threshold (step S12).

In step S12, when the correction operation type determination unit 23 determines that the amount of vertical movement of the work surface is greater than the certain threshold, the type of “movement in the vertical direction of the work surface” is used as the correction type. The type is determined by the type determination unit 23, and the correction type estimation process is terminated (step S19).

If the correction operation type determination unit 23 determines in step S12 that the amount of vertical movement of the work surface is equal to or less than the certain threshold value, the process proceeds to step S21, and in step S21, the current operation (operation) Whether the force bit is "1" or the suction force bit is "1" in the operation flag (validity flag) indicating that the progress information in the database 17 is "1") The type determination unit 23 makes the determination.

In step S21, if the force bit is “1” in the operation flag (the flag indicating validity) of the operation currently in progress (the progress information in the operation database 17 is “1”), the corrected operation type determination unit 23 In the case of the determination, the wiping operation is meant, so that the correction type is determined as “force correction” (step S17), and the correction type estimation process is terminated. On the other hand, if the bit of the suction force is “1” in the flag of the operation (the progress information in the operation database 17 is “1”) (the flag indicating the validity), the correction operation type determination unit 23 When the determination is made, it means that the cleaning is performed by suction. Therefore, it is determined that the correction type is “correction of suction force” (step S18), and the correction type estimation process is terminated.

If it is determined in step S3 that the operation is not performed in the operation database 17 by the correction operation type determination unit 23, the process proceeds to step S6. The correction operation type determination unit 23 determines whether the applied force is horizontal to the work surface and the amount of movement in the horizontal direction for a certain period of time is equal to or greater than a certain threshold value.

In step S6, when the force applied to the robot arm 5 by the human hand 16 is horizontal to the work surface and the amount of movement in the horizontal direction for a certain period of time is equal to or greater than the certain threshold value, the correction operation type determination unit 23 Is determined as the type of correction “region not desired to be worked” (step S8), and the correction type estimation process is terminated. In step S6, when the force applied to the robot arm 5 by the human hand 16 is not horizontal with respect to the work surface (for example, when it is vertical), or even when the force is horizontal to the work surface, the horizontal movement is performed. When the correction operation type determination unit 23 determines that the amount is less than the certain threshold, the correction type is determined as “no correction” (step S7), and the correction type estimation process ends.

As described above, the correction type can be switched by the correction operation type determination unit 23 without using the data input IF 26 such as a button.

The correction operation type determination unit 23 determines one type of the above seven types, but can also determine two types of correction at the same time.

The correction operation type determination method setting unit 27 in FIG. 3 sets the number of outputs determined by the correction operation type determination unit 23. However, the number of outputs may be input by the correction operation type determination unit 23 using the data input IF 26 and determined by a person.

The correction operation type determination unit 23 determines the correction type according to the number of outputs set by the correction operation type determination method setting unit 27. Specifically, when the number of outputs is 1, the correction type is determined by the algorithm of the correction type estimation method of FIG. 14, and when the number of outputs is “2”, the algorithm of FIG. To determine the type of correction. As a result, when the person who operates the household robot 1 is not familiar with the operation, setting the number of outputs to 1 makes it impossible to perform two types of correction at the same time, thus simplifying the operation. On the other hand, if you are used to the operation and want to correct two types at the same time, the correction can be performed efficiently by setting the number of outputs to a value of “2”.

The correction operation type determination unit 23 described above outputs one type, but as an example of outputting and correcting two types, the power of the operation database 17 is stronger during wiping work as shown in FIG. 18A. This is a case where it is necessary to clean the work surface at a higher speed than usual. In this case, the two types of wiping force and speed are corrected simultaneously. FIG. 18B shows a case where the suction force is increased while moving the work surface in parallel when cleaning is performed to suck dust and the like. In this case, the two types of movement and position of the work surface and suction force are corrected simultaneously.

The algorithm of the correction operation type determination unit 23 for executing the correction type estimation process of the correction operation type estimation method for outputting two types of types will be described in detail with reference to the flowchart of FIG.

As in the case of one type, when the power button 26a of the household robot 1 is turned “ON” and the robot arm 5 is gripped by the human hand 16 and no force is applied to the robot arm 5, The robot arm 5 does not move. When a force is applied to the robot arm 5 by the human hand 16, the robot arm 5 is moved in the direction in which the robot arm 5 is moved in the impedance control mode (a mode in which the force of the human hand 16 is detected and moved by impedance control). Can be made. In this case, the force detection unit 53 of the control unit 22 detects the force acting on the robot arm 5, and the information on the force detected by the force detection unit 53 is determined via the information acquisition unit 100. The data is input to the unit 23 (step S31).

Then, in step S32, all the components of the acquired in detected by the force detection unit 53 and the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is (specifically, _(f _dx of ID "1" in FIG. _{33, f dy, f dz,} f dφ, f dθ, f dψ)) is the threshold correction or not less whether the operation type determining unit 23 Judge with. All components were detected by the force detection unit 53 and acquired by the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is below the certain threshold If it is determined by the correction operation type determination unit 23, the robot arm 5 does not move and is not corrected (step S51), and the correction type estimation process of the correction operation type estimation method ends.

In step S32, one of the components of the acquired in detected by the force detection unit 53 and the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is the _{_{_{(specifically, (f dx, f dy,}}} f dz, f dφ, f dθ, f dψ) of ID in FIG. 33, "1") is a threshold determined by the correcting operation type determination unit 23 exceeds If so, the process proceeds to step S33.

In step S33, the corrective action type determination unit 23 further determines whether or not the current housework robot 1 is operating in the action database 17. Specifically, the cleaning operation is not selected by the operation selection unit 29, and the operation selection unit 29 does not select the operation for all the work IDs in the operation database 17, and the progress information is “0”. ”Is determined by the correction operation type determination unit 23 (when cleaning is not started), the correction operation type determination unit 23 determines that the operation database 17 is not operating, and the process proceeds to step S36. move on. When the operation selection unit 29 selects the cleaning operation and starts cleaning, and the correction operation type determination unit 23 determines that the progress information is “1”, the operation database 17 If it is operating, the correction operation type determination unit 23 determines that it is operating, and the process proceeds to step S34.

In step S34, when the robot arm 5 is gripped by the human hand 16 and a force is applied in the direction in which the cleaning operation of the robot arm 5 is to be corrected, the force applied to the robot arm 5 is detected by the force detector 53, power from the human hand 16 obtained by the detected and the information acquisition unit 100 by the force detection unit _{_{_{53 (f x, f y,}}} f z, f φ, f θ, f ψ) each having a certain time displacement of The amount is measured by the corrective action type determining unit 23, and the corrective action type is determined as to which of the position component (f _x , f _y , f _z ) or the posture component (f _φ , f _θ , f _ψ ) is larger. Measurement is performed by the unit 23. Specifically, as shown in FIG. 15, each time series force (f _x , f _y , f _z , f _φ , f _θ , f _ψ ) is measured by the correction operation type determination unit 23, and is constant. The correction operation type determination unit 23 measures how much the force is displaced in time (for example, time 1), and measures the component having the largest displacement. In this example, the displacement of f _phi is the largest, the process proceeds as judged by the correction operation type determining unit 23 orientation component is force is applied from the position component to step S39.

When the correction operation type determination unit 23 determines in step S34 that the displacement amount of the posture is larger than the displacement amount of the position, the correction operation type determination is made that the correction type is the type of “change in direction (posture)”. Then, the correction type estimation process is terminated (step S39).

On the other hand, when the correction operation type determination unit 23 determines in step S34 that the displacement amount of the position is equal to or greater than the displacement amount of the posture, the force component in the direction perpendicular to the work surface (for example, installed horizontally on the ground) In the case where the device 6 to be cleaned is f _z ), the correction operation type determination unit 23 determines whether or not the threshold value (specifically, f _dz of ID “1” in FIG. 33) is greater than or equal to (step S35). ). At the same time, a horizontal force component on the work surface (if cleaning the equipment 6 that is set horizontally, for example, ground f _x, either or both of the f _y) is, to a certain threshold value (specifically 33, the correction operation type determination unit 23 determines whether or not it is greater than or _{equal to} f _dx , f _dy of ID “1” in FIG. 33 (step S40).

If the correction operation type determination unit 23 determines in step S35 that the force component in the direction perpendicular to the work surface is smaller than the certain threshold value, it determines that there is no correction of the vertical surface (no type) and estimates the correction type. The process ends (step S45). If it is determined in step S40 that the force component in the direction parallel to the work surface is smaller than the certain threshold value, the correction operation type determination unit 23 determines that the horizontal plane is not corrected (no type), and correction type estimation processing is performed. Is finished (step S41).

If it is determined in step S40 that the correction action type determination unit 23 determines that the force component in the direction horizontal to the work surface is equal to or greater than the certain threshold value, the process proceeds to step S42.

In step S42, it is further determined whether or not the horizontal movement amount of the work surface is equal to or greater than a certain threshold (specifically, g _x , g _y of ID “2” in FIG. 33). Judge with. In step S42, the correction operation type determination unit 23 determines that the amount of horizontal movement of the work surface is greater than or equal to the certain threshold (specifically, g _x , g _y of ID “2” in FIG. 33). If it is determined, the correction operation type determination unit 23 determines that the correction type is “movement of the position of the work surface”, and ends the correction type estimation process (step S43).
If the correction operation type determination unit 23 determines in step S42 that the horizontal movement amount of the work surface is less than the certain threshold value, the type of “speed” in the direction horizontal to the work surface is used as the correction type. And the correction type estimation process is terminated (step S44).

If it is determined in step S35 that the corrective action type determining unit 23 determines that the force perpendicular to the work surface is equal to or greater than the certain threshold value, then whether or not the vertical movement amount of the work surface is greater than a certain threshold value. This is determined by the correction operation type determination unit 23 (step S46).

In step S46, when the correction operation type determination unit 23 determines that the vertical movement amount of the work surface is larger than the certain threshold value, the correction type is determined as the type of “movement in the vertical direction of the work surface”. Then, the correction type estimation process ends (step S50).

In step S46, when the correction operation type determination unit 23 determines that the vertical movement amount of the work surface is equal to or less than the certain threshold value, the process proceeds to step S47, and in step S47, the current operation is being performed (operation database). In the operation flag (the flag indicating effectiveness) of 17 progress information is “1”), the force bit is “1” in the case of the wiping operation, or the suction force bit is “1”. It is determined by the correction operation type determination unit 23.

In step S 47, if the force bit is “1” in the operation flag (validity flag indicating that the progress information in the operation database 17 is “1”), the correction operation type determination unit 23 In the case of the determination, it means a wiping operation, so that “correction of force” is determined as the correction type (step S48), and the correction type estimation process is terminated. On the other hand, if the bit of the suction force is “1” in the flag of the operation (the progress information in the operation database 17 is “1”) (the flag indicating the validity), the correction operation type determination unit 23 When the determination is made, it means that the cleaning is performed by suction. Therefore, it is determined that the correction type is “correction of suction force” (step S49), and the correction type estimation process is terminated.

If it is determined in step S33 that the operation is not performed in the operation database 17 by the correction operation type determination unit 23, the process proceeds to step S36. The correction operation type determination unit 23 determines whether the applied force is horizontal to the work surface and the amount of movement in the horizontal direction for a certain period of time is equal to or greater than a certain threshold value.

In step S36, the corrective action type determination unit 23 determines that the force applied to the robot arm 5 by the human hand 16 is horizontal to the work surface and the amount of horizontal movement for a certain period of time is equal to or greater than a certain threshold value. In this case, it is determined that the type of correction is “region not desired to be worked” (step S38), and the correction type estimation process is terminated. In step S36, when the force applied to the robot arm 5 by the human hand 16 is not horizontal with respect to the work surface (for example, when it is vertical), or even when the force is horizontal to the work surface, the movement in the horizontal direction is performed. When the correction operation type determination unit 23 determines that the amount is less than the certain threshold, the correction type is determined as “no correction” (step S36), and the correction type estimation process ends.

As described above, two or more types of correction can be switched by the correction operation type determination unit 23 without using the data input IF 26 such as a button.

The motion correction unit 20 is a function that corrects motion information in the motion database 17 by applying a force to the robot arm 5 with a human hand 16 during motion based on the position, posture, and time of the motion database 17. .

Hereinafter, the function of the motion correction unit 20 will be described.

When the power is turned on by a human hand 16 using a data input IF 26 (for example, the power button 26a of the operation panel 26A) disposed on the work table 7 of the housework robot 1, the operation correction unit 20 operates in the impedance control mode. A command is issued to the control parameter management unit 21.

Next, with the human hand 16, the operation selection unit 29 selects a desired work from the work list in the operation database 17 and gives an instruction to start the operation. The motion correction unit 20 is based on the motion information of the work ID selected from the motion database 17 (specifically, the position of the rail movable unit 8b and the position, posture, and time of the robot arm 5). The control parameter management unit 21 is instructed to operate the robot arm 5 in the force hybrid impedance control mode.

In the case of the force hybrid impedance control mode, among the flags for the action IDs in the action database 17 (flags indicating validity), for each position and posture of the robot arm 5 in which the flag bit is “1”. In addition, a hybrid impedance control mode (a mode in which the robot arm 5 operates according to a force applied to the robot arm 5 by a person or the like while operating in the position control mode) is set by the operation correction unit 20, and suction force is set. Alternatively, a force control mode is set in the operation correction unit 20 for a component in which a bit of a force flag (a flag indicating effectiveness) is “1”. Of the six components of position and orientation, the component for which neither the hybrid impedance control mode nor the force control mode is set has the impedance control mode set by the operation correction unit 20. For example, when the work ID in FIG. 4 is “1”, the work for cleaning by sucking dust is shown, and the flag when the work ID is “1” and the operation ID is “1” is 1. , 2 and 14th bits are “1”, so that the hybrid impedance control mode is set by the operation correction unit 20 for the x-axis and y-axis components, and the force control mode for the z-axis components. Is set by the motion correction unit 20, and the impedance control mode is set by the motion correction unit 20 for the posture component. When the work ID in FIG. 4 is “2”, it indicates a work for wiping and cleaning. When the work ID is “2” and the operation ID is “1”, only the first, second, and eighth bits are “ 1 ”, the hybrid impedance control mode is set by the motion correction unit 20 for the x-axis and y-axis components, the force control mode is set by the motion correction unit 20 for the z-axis component, and impedance control is performed for the posture component. The mode is set by the operation correction unit 20.

The control parameter management unit 21 receives a command from the operation correction unit 20. That is, when a command is issued from the motion correction unit 20 to the control parameter management unit 21 so as to perform the cleaning work in the force hybrid impedance control mode, the housework is performed at the position commanded by the rail movable unit 8b as shown in FIGS. 16A to 16C. While the robot 1 is self-propelled, the robot arm 5 starts the cleaning operation with the position, posture and force or suction force of the operation ID.

Next, an example will be described in which a person confirms the state of dirt on the work surface and the like, and as shown in FIG. 12C, the robot arm 5 is moved a little further in the lateral direction to be cleaned (sucked).

As shown in FIG. 12C, the robot arm 5 is directly gripped by a human hand 16 and a force is applied to the robot arm 5 parallel to the work surface so as to move in parallel with the work surface.

The correction type estimation shown in the flowchart of FIG. 14 based on the force applied by the human hand 16 to the robot arm 5 and the information stored in the operation database 17 acquired by the information acquisition unit 100 by the correction operation type determination unit 23. The type of correction is estimated and determined by processing. Here, since the human hand 16 applies a force in a direction parallel to the work surface to the robot arm 5 to move the robot arm 5 by a certain threshold value or more, in step S14, the correction type is “the position of the work surface. The correction operation type determination unit 23 determines that the type is “movement”.

When the work ID in FIG. 4 is “1” and the work ID is “1”, the x-axis component and the y-axis component move the robot arm 5 in the position control mode by the force hybrid impedance control mode. However, the force applied to the robot arm 5 by the human hand 16 in the impedance control mode is detected by the force detection unit 53, and the robot arm 5 is moved in the x-axis direction in the direction in which the force is applied to the robot arm 5 by the human hand 16. The cleaning position can be corrected as shown in FIG. 12D by moving in the direction and the y-axis direction.

In this example, since it is desired to correct the operation only in the x-axis direction and the y-axis direction, the correction operation type determination unit 23 performs the correction operation type determination unit 23 at the timing when the correction type is determined. By setting the 0th and 1st bits of the correction parameter flag to “1” and setting the other bits to “0” and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, It can be set so that movements other than the direction of the x-axis and the direction of the y-axis cannot be performed. Furthermore, by changing the mechanical impedance setting value in the impedance control mode by the correction operation type determination unit 23 and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, the x-axis direction and y of the robot arm 5 The rigidity in the axial direction is made lower than the rigidity in the other directions so that the robot arm 5 can be easily moved in the x-axis direction and the y-axis direction by the human hand 16, and the rigidity in directions other than the x-axis direction and the y-axis direction can be increased. The robot arm 5 can be made difficult to move in a direction other than the x-axis direction and the y-axis direction by the human hand 16. Thereby, when it is desired to correct only the x-axis component and the y-axis component of the robot arm 5, it is possible to prevent the z-axis component of the robot arm 5 from being erroneously corrected. Further, during the correction of the robot arm 5 in the x-axis direction and the y-axis direction, the correction operation type determination unit 23 causes the suction strength of the z-axis component or the force applied to the work surface to be weaker than the operation before the correction or It can also be made smaller (specifically about half). Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the suction or force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. As a result, during the correction by moving the robot arm 5 in the x-axis direction and the y-axis direction, force is applied to the robot arm 5 to damage the device 6 or accidentally suck in anything other than dust. Can be prevented.

As described above, the robot arm 5 is grasped by the human hand 16 and the robot arm 5 is moved in the x-axis direction and the y-axis direction by Δx and Δy by applying a force in a direction parallel to the work surface. In this case, the value of Δx and the value of Δy are transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 subtracts Δx from all the x coordinate values of the motion information of the selected work ID, and further subtracts Δy from all the y coordinate values to correct the motion information corrected by the motion correction unit 20. To the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates for Δx and Δy. Thereby, it correct | amends to operation | movement like FIG. 12D. Next, the operation information reduced by Δx and Δy is stored in the operation database 17 by the operation storage unit 15.

Next, as shown in FIG. 32B, for example, when working on the protruding portion 6 a installed on the device 6 while working on the device 6, the robot arm 5 is directly gripped by a human hand 16, A force is applied to the robot arm 5 perpendicular to the work surface so as to move in a direction perpendicular to the robot arm 5.

In the correction type estimation process shown in the flowchart of FIG. 14 based on the force applied to the robot arm 5 by the human hand 16 and the information in the action database 17 respectively acquired by the information acquisition part 100 by the correction action type determination part 23. The type of correction is estimated and determined. Here, since the robot arm 5 is moved by a human hand 16 in a direction perpendicular to the work surface by moving the robot arm 5 by a threshold value or more, in step S19, the correction type is “vertical direction of the work surface”. Is determined by the correction operation type determination unit 23.

While the robot arm 5 is moved in the position control mode by the force hybrid impedance control mode, the force of the human hand 16 is detected by the force detection unit 53 by the impedance control mode and the force is applied to the robot arm 5 by the human hand 16. The cleaning position can be corrected as shown in FIG. 32C by moving the robot arm 5 in the z-axis direction in the applied direction.

In this example, since it is desired to correct the operation only in the z-axis direction, the correction operation type determination unit 23 performs the second bit of FIG. 6 at the timing when the correction type is determined by the correction operation type determination unit 23. By setting “1” and setting the other bits to “0” and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, movement other than in the z-axis direction cannot be performed. Can be set. Further, by changing the mechanical impedance set value in the impedance control mode by the correction operation type determination unit 23 and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, the rigidity in the z-axis direction is changed to another direction. The rigidity of the robot arm 5 is made easier to move in the z-axis direction with the human hand 16 and the rigidity other than in the z-axis direction is increased, and the robot arm 5 is moved in the direction other than the z-axis direction with the human hand 16. It can be made difficult to move in the direction.

Further, when correcting the movement of the robot arm 5 in the z-axis direction, the correction action type determination unit 23 causes the suction strength of the z-axis component or the force applied to the work surface to be weaker or smaller than that before the correction. (Specifically, about half). Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the suction or force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. This prevents the robot arm 5 from being damaged by applying force or suction to the robot arm 5 while moving the robot arm 5 in the z-axis direction, or accidentally sucking in anything other than dust. it can.

As described above, when the robot arm 5 is gripped by a human hand 16 and a force is applied in a direction perpendicular to the work surface to move the robot arm 5 in the z-axis direction by Δz, the value of Δz is Then, the data is transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 transmits motion information corrected by subtracting Δz from all z coordinate values of the motion information of the selected work ID from the motion correction unit 20 to the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates corresponding to Δz. This corrects the operation as shown in FIG. 32C. Next, the motion information reduced by Δz is stored in the motion database 17 by the motion storage unit 15.

As shown in FIG. 30B, for example, when the suction nozzle 9 or the mop 10 of the robot arm 5 is changed in length and cleaned, the robot arm 5 is directly gripped by a human hand 16 as shown in FIG. 30C. The robot arm 5 is moved in the direction in which the longitudinal direction is desired to be changed.

In the correction type estimation process shown in the flowchart of FIG. 14 based on the force applied to the robot arm 5 by the human hand 16 and the information in the action database 17 respectively acquired by the information acquisition part 100 by the correction action type determination part 23. The type of correction is estimated and determined. Here, since force is applied to the robot arm 5 to move the longitudinal direction of the

cleaning units

9 and 10 in the direction to be changed by the human hand 16, in step S9, the correction type “direction (posture)” is selected. The correction operation type determination unit 23 determines that the type is “change”.

While the robot arm 5 is moved in the position control mode by the force hybrid impedance control mode, the force applied to the robot arm 5 by the human hand 16 is detected by the force detection unit 53 by the impedance control mode. The cleaning direction can be corrected as shown in FIG. 30D by rotating the robot arm 5 in the φ axis direction in the direction in which force is applied to the robot arm 5.
In this example, since it is desired to correct the operation only in the direction of the φ axis, the correction operation type determination unit 23 sets the correction parameter flag in FIG. 6 at the timing when the correction type is determined by the correction operation type determination unit 23. The third bit is set to “1” and the other bits are set to “0”, and a command is issued from the correction operation type determination unit 23 to the control parameter management unit 21. Accordingly, the correction operation type determination unit 23 can set the movement other than the direction of the φ axis so that the movement is not possible. Further, the mechanical impedance set value in the impedance control mode is changed by the correction operation type determination unit 23, and a command is issued from the correction operation type determination unit 23 to the control parameter management unit 21, so that the rigidity in the φ-axis direction is changed to another direction. The rigidity of the robot arm 5 is made easier to move in the φ-axis direction by the human hand 16 and the rigidity other than the φ-axis direction is increased, and the robot arm 5 is moved in the direction other than the φ-axis direction by the human hand 16. It can be made difficult to move.

Further, during the correction of the robot arm 5 in the φ-axis direction, the correction operation type determination unit 23 causes the strength of suction of the z-axis component or the force applied to the work surface to be weaker or smaller than the operation before the correction (specifically Can be halved). Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the suction or force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. Thereby, it is possible to prevent the device 6 from being damaged by applying a force to the robot arm 5 while moving in the φ-axis direction, or accidentally sucking anything other than dust.

As described above, when the robot arm 5 is gripped by a human hand 16 and a force is applied in a direction perpendicular to the work surface to rotate the robot arm 5 in the φ axis direction by Δφ, the value of Δφ is Then, the data is transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 transmits the motion information corrected by subtracting Δφ from all the φ coordinate values of the motion information of the selected work ID from the motion correction unit 20 to the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates corresponding to Δφ. This corrects the operation as shown in FIG. 30D. Next, the operation information reduced by Δφ is stored in the operation database 17 in the operation storage unit 15.

As described above, the motion correction unit 20 is generated by applying force to the robot arm 5 with the human hand 16 while operating in the force hybrid impedance control mode according to the position, posture, and time of the motion database 17. The corrected position can be corrected for each direction.

Next, as shown in FIG. 27B, when changing the force on the work surface at the time of wiping and cleaning, the robot arm 5 is directly gripped by a human hand 16 and the force is applied to the robot arm 5 in a direction perpendicular to the work surface. Call.

In the correction type estimation process shown in the flowchart of FIG. 14 based on the force applied to the robot arm 5 by the human hand 16 and the information in the action database 17 respectively acquired by the information acquisition part 100 by the correction action type determination part 23. The type of correction is estimated and determined. Here, since force is applied to the robot arm 5 in the direction perpendicular to the work surface by the human hand 16 and the robot arm 5 is not moved beyond a certain threshold, the type of “force correction” is selected as the type of correction in step S17. Is determined by the correction operation type determination unit 23.

At the timing when the correction operation type determination unit 23 determines that the correction type is “force correction”, the correction operation type determination unit 23 controls the control parameter management unit to operate in the high-rigidity position control mode from the force hybrid impedance control mode. A command is issued to 21. In the high-rigidity position control mode, the high-rigidity at the time of position control can be set in the high-rigidity position control mode by the correction-operation-type determining unit 23 at the time of command from the correction operation type determining unit 23 to the control parameter management unit 21. Since the operation flag of operation ID “2” and operation ID “1” in the operation database 17 is set to “1” in the 0, 1, and 8 bits, the z-axis direction operates in the force control mode. Since the other directions operate in the hybrid impedance control mode, the control parameter management unit 23 controls the control parameters so that only the z-axis direction operates in the high-rigidity position control mode and the other directions operate in the hybrid impedance control mode. A command is issued to the unit 21.

Next, as shown in FIG. 27B, while the robot arm 5 is being wiped and cleaned, the robot arm 5 is directly gripped by a human hand 16 while the robot arm 5 is operating in a heavily contaminated part. When it is desired to wipe the surface with a strong surface, a human hand 16 applies a downward force to the robot arm 5 (for example, the mop 10 of the robot arm 5) toward the work surface. The high-rigidity position control mode is a mode in which the position control mode set for each direction among the hybrid impedance control modes at the time of cleaning is further increased in rigidity, and the gain in the position error compensation unit 56 is increased (specifically In this case, the robot arm 5 can be prevented from being easily moved by applying force to the robot arm 5 with a human hand 16. The force applied to the robot arm 5 by the human hand 16 can be detected by the force detection unit 53. The force detected by the force detection unit 53 of the control unit 22 is notified to the operation correction unit 20. By storing the force notified to the motion correction unit 20 in the motion database 17 in the motion storage unit 15, the motion can be corrected so as to strongly wipe only the dirty part.

When the person wants to finish the correction, the robot arm 5 is gripped to stop applying force to the robot arm 5.

When no force is applied to the robot arm 5 with the human hand 16, all the components of the force fall below the threshold value in step S2 in FIG. "None" (step S20 in FIG. 14). The motion correction unit 20 receives the information of “no correction” and issues a command from the correction operation type determination unit 23 to the control parameter management unit 21 so as to control from the highly rigid position control mode to the hybrid impedance control mode. As a result, the corrected operation database 17 is cleaned.

As described above, the motion correction unit 20 uses the force information in the motion database 17 to perform cleaning with the corrected force when the human hand 16 applies force while operating in the hybrid impedance control mode. It becomes possible to correct.

Next, as shown in FIG. 28C, when the suction force to the work surface is changed, the robot arm 5 is directly gripped by a human hand 16 and a force is applied to the robot arm 5 in a direction perpendicular to the work surface.

In the correction type estimation process shown in the flowchart of FIG. 14 based on the force applied to the robot arm 5 by the human hand 16 and the information in the action database 17 respectively acquired by the correction action type determination part 23 by the information acquisition part 100. The type of correction is estimated and determined. Here, since a force is applied to the robot arm 5 by the human hand 16 in a direction perpendicular to the work surface and the robot arm 5 is not moved beyond a certain threshold value, in step S18, the correction type is “correction of suction force”. Is determined by the correction operation type determination unit 23.

Control parameter management from the correction operation type determination unit 23 to operate from the force hybrid impedance control mode to the high-rigidity position control mode at the timing determined by the correction operation type determination unit 23 as “correction of attraction force” as the correction type. A command is issued to the unit 21. In the high-rigidity position control mode, the high-rigidity at the time of position control can be set in the high-rigidity position control mode by the correction-operation-type determining unit 23 at the time of command from the correction operation type determining unit 23 to the control parameter management unit 21. Since the operation flag of the operation database 17 with the operation ID “1” and the operation ID “1” is set to “1” in the 0, 1, and 14 bits, the z-axis direction operates in the suction control mode. Since the other directions operate in the hybrid impedance control mode, the control parameter management unit 23 controls the control parameters so that only the z-axis direction operates in the high-rigidity position control mode and the other directions operate in the hybrid impedance control mode. A command is issued to the unit 21.

Next, as shown in FIG. 28B, while the robot arm 5 is being suction-cleaned, the robot arm 5 is directly gripped by a human hand 16 while the robot arm 5 is operating in a heavily contaminated part, When it is desired to suck the work surface strongly, a human hand 16 applies a downward force to the robot arm 5 (for example, the mop 10 of the robot arm 5) toward the work surface. The high-rigidity position control mode is a mode in which the normal position control mode is further increased in rigidity, and is realized by increasing the gain in the position error compensation unit 56, and force is applied to the robot arm 5 with the human hand 16. When applied, the robot arm 5 can be prevented from moving easily, and the force applied to the robot arm 5 by the human hand 16 can be detected by the force detection unit 53. The force detected by the force detection unit 53 of the control unit 22 is notified to the motion correction unit 20 via the control parameter management unit 21, and the motion correction unit 20 uses the suction force in the z-axis direction of the motion database 17 as the motion database. Using the conversion table shown in FIG. 19 stored in 17 (or the storage unit of the operation correction unit 20), the force is converted into the suction force. For example, when the force applied to the robot arm 5 by the person is 4.5 [N], the suction force is converted to “4” corresponding to the force 4 to 5 [N] from the conversion table. By storing the suction force “4” in the operation database 17 in the operation storage unit 15, it is possible to correct the operation so that only the dirty portion is strongly suctioned and cleaned. When a person wants to end the correction, the robot arm 5 may be held by the person's hand 16 and the force applied to the robot arm 5 by the person's hand 16 may be stopped. That is, when no force is applied to the robot arm 5 with the human hand 16, all components of the force are equal to or less than the threshold values in step S <b> 2 of FIG. 14, so that the correction operation type determination unit 23 sets “ "No correction" is determined (step S20 in FIG. 14). In response to the determination that the correction type is “no correction”, the motion correction unit 20 issues a command to the control parameter management unit 21 to control from the high-rigidity position control mode in the hybrid impedance control mode. As a result, the corrected operation database 17 is cleaned.

As described above, the motion correction unit 20 applies the force to the robot arm 5 with the human hand 16 while operating in the hybrid impedance control mode by the suction force of the motion database 17, thereby correcting the suction force. It can be corrected to clean.

Next, as shown in FIG. 29D, when the cleaning speed is changed, the robot arm 5 is directly held by the person's hand 16 and accelerated. In order to apply a force to the robot arm 5 and decelerate, the force is applied to the robot arm 5 with a human hand 16 in the direction opposite to the direction of cleaning. At this time, the speed of the hand position of the robot arm 5 may be changed, but the force is applied to the robot arm 5 with the human hand 16 so as not to move the position beyond the certain threshold.

In the correction type estimation process shown in the flowchart of FIG. 14 based on the force applied to the robot arm 5 by the human hand 16 and the information in the action database 17 respectively acquired by the correction action type determination part 23 by the information acquisition part 100. The type of correction is estimated and determined. Here, since a force is applied to the robot arm 5 in the direction horizontal to the work surface with the human hand 16 and the robot arm 5 is not moved more than the certain threshold value, the work surface is selected as the correction type in step S15 of FIG. The correction operation type determination unit 23 determines that the type is “speed” in the horizontal direction.

While the robot arm 5 is moved in the position control mode by the hybrid impedance control mode, the force applied to the robot arm 5 by the human hand 16 is detected by the force detection unit 53 by the impedance control mode, and the robot is moved by the human hand 16. The robot arm 5 is moved in the x-axis direction and the y-axis direction in the direction in which a force is applied to the arm 5. From the position of the robot arm 5 shown in the operation database 17, for example, work ID operation ID as _{_{_{(x 1, y 2, z}}} 1), the position of the robot arm 5 of the next operation _{_{ID (x 2, y 2,}} z when the time to move to ₂₎ and _{t 1,} the case of changing the speed of the robot arm 5 by the force of the human hand 16 (see FIG. 29C), i.e., the position _(x _{1, y} 2, z ₁ ) to the position (x ₂ , y ₂ , z ₂ ), when the time taken to move from t ₁ to t ₂ is changed, the value of time t ₂ is the control unit 22 and control parameter management. The data is transmitted to the operation correction unit 20 via the unit 21. The motion correction unit 20 changes the motion information of the selected work ID from the time t _{1 to} the time t ₂ and transmits the motion information from the motion correction unit 20 to the control parameter management unit 21. In the control parameter managing unit 21 instructs the control parameter managing unit 21 to operate in t ₂ is corrected time to the control unit 22.

This corrects the operation as shown in FIG. 29D. Next, the time t ₂ is stored in the operation database 17 by the operation storage unit 15.

As described above, the motion correction unit 20 applies the force to the robot arm 5 with the human hand 16 while operating in the force hybrid impedance control mode based on the information on the position, posture, and time of the motion database 17. The operation speed of the robot arm 5 can be corrected.

As shown in FIG. 31, a case where an area RB that the housework robot 1 does not want to work is set using the robot arm 5 will be described as an example.

When the power is turned on by a human hand 16 using a data input IF 26 (for example, the power button 26a of the operation panel 26A) disposed on the upper part of the housework robot 1, the operation correction unit 20 is controlled to operate in the impedance control mode. A command is issued to the parameter management unit 21. In a state where the operation is not selected by the motion selection unit 29, as shown in FIG. 31, the hand 16 of the person 16A directly grips the robot arm 5 (or the cleaning units 9 and 10) and touches the work surface. The robot arm 5 is moved so as to move in parallel, and the robot arm 5 is moved along the outline of the region RB that is not desired to be worked. FIG. 20A is a view of the work surface as viewed from above. If a region RB that is not desired to be worked is a hatched region, the human hand 16 moves the robot arm 5 (or the cleaning unit 9 or 10), As indicated by the arrow, the robot arm 5 (or the cleaning units 9 and 10) is moved along the outline of the region RB that is not desired to be worked. At that time, a mark 63 is given to the center tip of the upper surface of the suction nozzle 9 (or the mop 10) attached to the hand (hand 30) of the robot arm 5 (see FIGS. 31, 20A and 20B). And move the mark 63 in the direction that you do not want.

The correction operation type determination unit 23 executes the correction type estimation process shown in FIG. 14 to determine that the operation database 17 is not operating (steps S2, S3, and S6). If the force applied to 5 is horizontal to the work surface and the amount of movement in the horizontal direction for a certain period of time is equal to or greater than the certain threshold value, in step S8, the “type of region that you do not want to work” is selected as the correction type. It is determined that it is a type.

In the impedance control mode, the force applied to the robot arm 5 by the human hand 16 is detected by the force detection unit 53, and the robot arm 5 is moved in the x-axis direction and in the direction in which the force is applied to the robot arm 5 by the human hand 16. As shown in FIG. 20A, the position (x ₁ , y ₁ ), the position (x ₂ , y ₂ ), the position (x ₃ , y ₃ ), and the position (x ₄ , y ₄ ) are moved in the y-axis direction. When the suction nozzle 9 of the robot arm 5 is sequentially moved, these pieces of position information are transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21. In response to the command, the motion correction unit 20 stores these pieces of position information as information on the work impossible region RB in the work unavailable region database 28 by the operation storage unit 15. The fact that these four positions are information on the apex of the work impossible area RB means that, for example, the hand position of the robot arm 5 moved by a person at a certain interval is acquired and the coordinates of the acquired hand position are connected. An area can be generated and used as an unworkable area RB. In the correction operation type determination method setting unit 27, a function for determining the shape of the region is added. For example, when “rectangular” is set, the moving direction is changed at an angle close to 90 degrees. For example, the position is stored as vertex information, and when “random” is set, the hand position of the robot arm 5 moved by a person at a certain interval is acquired, and the coordinates of the acquired hand position are connected. And can be used as an unworkable area RB.

In this example, since it is desired to correct the motion of the robot arm 5 only in the x-axis direction and the y-axis direction, the correction parameter shown in FIG. 6 is displayed at the timing when the correction type is determined by the correction operation type determination unit 23. The robot arm 5 is set by setting the 0th and 1st bits of the flag to “1” and setting the other bits to “0” and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21. Can be set so as not to move in an axial direction other than the x-axis direction and the y-axis direction. Furthermore, by changing the mechanical impedance setting value in the impedance control mode and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, the rigidity in the x-axis direction and the y-axis direction can be reduced, The hand 16 makes it easy to move the robot arm 5 in the x-axis direction and the y-axis direction, and increases the rigidity in the axial direction other than the x-axis direction and the y-axis direction. It can be made difficult to move in directions other than the y-axis direction.

As described above, the motion correction unit 20 can set an area where the user's hand 16 does not want to perform work (for example, cleaning work) by applying force.

As shown in FIG. 21, the display unit 14 displays the screen of the display unit 14 divided into left and right two

screens

14a and 14b, and the operation of the robot arm 5 described in the operation database 17 is displayed on the left screen 14a. Is displayed as video, photo or text. Further, the correction type information estimated by the correction operation type determination unit 23 is displayed on the screen 14b on the right side as video, photos, or text. In the example of FIG. 21, when a human hand 16 applies a force perpendicular to the work surface to the robot arm 5 and performs an operation to correct the force application, the correction operation type determination unit 23 sets “ At the timing determined to be the type of “force correction”, the image on which the force is corrected and the strength of the current force are displayed on the right screen 14b.

In this example, video, photo, or text is used, but audio that explains the operation may be used.

The operation steps of the operation correction unit 20, the correction operation type determination unit 23, the operation selection unit 29, the cleaning operation storage unit 15, the operation database 17, and the control parameter management unit 21 (that is, cleaning after the housework robot 1 starts to be driven). Cleaning work and cleaning operation setting processing performed until work is started will be described based on the flowchart of FIG.

The power of the household robot 1 is turned on by the human hand 16 using the data input IF 26 (step S121).

Next, the operation correction unit 20 issues a command to the control parameter management unit 21 to control in the impedance control mode (step S122).

Next, the correction operation type determination unit 23 determines whether or not the work impossible area RB is to be corrected (step S130). If the correction operation type determination unit 23 determines that the work impossible area RB is corrected, the operation correction unit 20 performs correction (step S133), and the correction information is stored in the operation database 17 by the operation storage unit 15. Store (step S134). Thereafter, the process proceeds to step S123.

If it is determined in step S130 that the correction operation type determination unit 23 does not correct the work impossible area RB, or if step S134 is executed, the process proceeds to step S123. In step S123, the operation selecting unit 29 selects one cleaning operation from the cleaning operation list displayed on the display unit 14 through the data input IF 26, and the operation selection unit 29 selects the progress information in the operation database 17 as progress information. The current cleaning work is set (step S123).

Next, the motion correction unit 20 instructs the control parameter management unit 21 to operate in the force hybrid impedance control mode, and guides the robot arm 5 to the work surface such as the device 6 with a human hand 16 to input the data input IF 26 ( For example, an instruction to start the cleaning operation is issued using the start button of the cleaning switch 26c (step S124).

Next, when a person applies a force in the direction to be corrected, the correction operation type determination unit 23 estimates and determines the type of the correction operation (step S125).

Next, in step S125, when the correction operation type determination unit 23 determines that the type of correction is the type of force applied to the work surface or the type of suction force, highly rigid position control in the direction perpendicular to the work surface. A command is issued from the operation correction unit 20 to the control parameter management unit 21 so as to operate in the mode (steps S126 and S127).

Next, the motion correction unit 20 corrects the motion information by holding the robot arm 5 with the human hand 16 and applying a force to the robot arm 5 with the human hand 16 in the direction to be corrected (step S128).

On the other hand, if it is determined in step S125 that the correction type is a correction type other than the force applied to the work surface or the suction force type, the control mode is not changed in the force impedance control mode, and the person in the direction to be corrected is selected. By applying force to the robot arm 5 with the hand 16, the motion correction unit 20 corrects the motion information (steps S 126 and S 128).

Next, the cleaning operation information corrected in step S128 is stored in the operation database 17 by the operation storage unit 15, and the series of cleaning work and cleaning operation setting processing ends (step S129).

On the other hand, if the correction operation type determination unit 23 determines in step S125 that the correction type is “no correction”, the series of cleaning work and cleaning operation setting processing ends (steps S126 and S131).

After the setting process of the cleaning work and cleaning operation is completed, the housework robot 1 performs cleaning based on the set cleaning work and cleaning operation.

By the above operation steps S121 to S122, step S130, step S123, step S124, step S133 to step S134, step S125 to step S128, and step S131, the hybrid impedance control mode or the high rigidity can be obtained during the operation with the force hybrid impedance control. The cleaning operation by the robot arm 5 is realized by correcting the cleaning operation by the position control.

Further, the correction operation type determination unit 23 can automatically switch and correct a plurality of cleaning operations by simply applying force to the robot arm 5 with a human hand 16 without using a button or the like.

Furthermore, the correction operation type determination unit 27 allows a person who is accustomed to the operation of the robot arm 5 or a skilled person to perform two types of correction at the same time at one time. Types of corrections can be made.

Further, by including the control parameter management unit 21 and the control unit 22, the machine impedance value of the robot arm 5 is appropriately set according to the type of the correction operation, so that the machine arm according to the correction direction of the robot arm 5. Since it can be controlled by changing the impedance value, or the suction force or force being corrected can be weakened or stopped, the device 6 may be damaged or a thing other than dust may be accidentally sucked in during the correction of the cleaning operation. Can be prevented.

In the first embodiment, the motion correction unit 20 includes the force applied to the robot arm 5 by the human hand 16 and the information in the motion database 17 respectively acquired by the information acquisition unit 100 by the correction operation type determination unit 23. After the estimation of the type of correction, the cleaning operation was corrected immediately, but the human arm 16 accidentally applied force to the robot arm 5 to select the type of correction that was not intended by the person. In order to prevent this, the correction operation type determination unit 23 may start the correction after a certain time after estimation. In this case, until the correction is started, the person can operate as many times as necessary until the intended correction type is selected.

In the first embodiment, the operation selection unit 29, the operation storage unit 15, the operation correction unit 20, the correction operation type determination unit 23, the correction operation type determination method setting unit 27, the control parameter management unit 21, and the control unit 22 Etc., or any part of them, can itself be composed of software. Therefore, for example, as a computer program having steps constituting the control operation of the first embodiment of the present specification or an embodiment to be described later, it is readable and stored in a recording medium such as a storage device (hard disk or the like), and the computer Each step described above can be executed by reading the program into a temporary storage device (semiconductor memory or the like) of the computer and executing it using the CPU.

(Second Embodiment)
In the second embodiment of the present invention, the basic configuration of the robot arm control device of the housework robot 1 including the robot arm control device is the same as that of the first embodiment, so the description of the common parts is as follows. Omitted and only different parts will be described in detail below.

In the second embodiment, as shown in FIG. 34, the housework robot 1 is a robot that uses a robot arm 5 to stir the ingredients in the pot 3. That is, the housework robot 1 according to the second embodiment uses the robot arm 5 in the home to perform housework work while acting on the object of housework (cooking work) (for example, ingredients in the pot 3). Robot to perform.

As shown in FIG. 34, the robot arm 5 of the household robot 1 is installed on the wall surface 7a of the work table 7, and the base end of the robot arm 5 is movably supported by the rail 8 fixed to the wall surface 7a. The robot arm 5 can be moved on the rail 8 in the lateral direction along the rail 8, for example, in the horizontal direction, automatically by the force of the hand 16 of the person 16A, or by a motor or the like. The fixing position of the base end of the robot arm 5 is not limited to the wall surface 7a of the work table 7, but may be a ceiling or the like.

The rail 8 has a rail fixed portion 8a fixed to the wall surface 7a and a rail movable portion that has a wheel (not shown) that is driven to rotate forward and backward by driving of the motor 65 and is movable with respect to the rail fixed portion 8a. 8b. The base part 34 to which the base end of the robot arm 5 is connected is connected to the rail movable part 8b, and the base part 34 of the robot arm 5 is configured to be movable together with the rail movable part 8b with respect to the rail fixing part 8a. ing. Alternatively, instead of such a configuration, a wheel that is driven to rotate forward and backward by a motor 65 is provided on the base portion 34 to which the base end of the robot arm 5 is connected, and along the rail 8 fixed to the wall surface 7a, It is good also as a structure that the base part 34 moves.

At the tip of the robot arm 5 is attached a hand 30 that can detachably hold the ladle 4 as an example of a cooking utensil for stirring the pot 3.

First, as in the first embodiment, the data input IF 26 (for example, pressing “ON” of the power button 26a of the operation panel 26A in FIG. 26) in which the hand 16 of the person 16A is arranged on the side surface of the work table 7 is used. Turn on the power. Note that the operation of detachably holding the ladle 4 on the hand 30 is the same as that of attaching the suction nozzle 9 or the mop 10 to the hand in the first embodiment, and thus the description thereof is omitted.

Next, when the data input IF 26 (for example, the start button of the switch 26c of the operation panel 26A in FIG. 26) disposed on the side surface of the work table 7 is pushed with the hand 16 of the person 16A, the housework robot 1 is activated, An operation selecting unit 29 described later selects an optimal operation, for example, a housework operation (cooking operation), and starts a housework operation (cooking operation) based on the selected operation. Specifically, as shown in FIG. 36A, the ladle 4 attached to the hand 30 of the robot arm 5 starts the operation of stirring the whole ingredients in the pot while rubbing the pot bottom in the pot 3.

Next, the person 16A confirms the condition of the ingredients of the pan 3, and corrects the operation of the robot arm 5 with a strong force so that the ladle 4 is mixed while rubbing the pan bottom strongly as shown in FIG. 36B. In this way, as shown in FIG. 36C, the entire ingredients can be mixed while rubbing the pan bottom with a strong force in the ladle 4 held by the robot arm 5.

FIG. 43 is a diagram showing in detail the components of the control device for the robot arm 5 constituting the housework robot 1, and similarly to the first embodiment, the control device main body 45 and the motion generation device 12 for generating motion. FIG. 4 is a diagram illustrating a detailed configuration of a robot arm 5 to be controlled, a rail 8, and a peripheral device 47. The control device of the housework robot 1 is roughly composed of a control device main body 45, the motion generation device 12, and a peripheral device 47. Since the control device main body 45 and the robot arm 5 are the same as those in the first embodiment, description thereof is omitted.

The motion generation device 12 includes a motion database 17, an unworkable area database 28, a corrected motion type determination method setting unit 27, a motion correction unit 20, a corrected motion type determination unit 23, and an object state determination unit (object An example of a state determination unit) 19, an operation storage unit 15, an operation selection unit 29, and an information acquisition unit 100 are provided.

As shown in FIG. 44A as an example, the work procedure information database 18 identifies a work ID number for identifying a work, a process ID number for identifying individual processes of the work, and an operation of the robot arm 5 in the process. Operation ID number, information related to the process corresponding to the process ID number, elapsed time that is the time taken for the process, amount to be cooked in the process, and state ID number that identifies the state of the object in the process And progress information indicating whether or not it is currently operating. Here, the state of the object means an object when cooking work is performed on an object (for example, ingredients in the pan 3) of housework (for example, cooking work) using the robot arm 5 in the home. (For example, ingredients in the pan 3) means information on the state (for example, a viscous state, a rigid state, and a burned state, which will be described later).

The status ID indicates the status information shown in FIG. 44C as an example. In FIG. 44C, the state ID includes a state ID number that identifies the state, an object information ID number that is an ID number that identifies object information (object information) in the process, and the object information ID. The information indicating the viscosity of the object information indicated by the object, the information indicating the rigidity indicated by the object information ID, the information indicating the degree of burning of the object information ID, and the viscosity, rigidity, and burning Priority state information indicating whether the state is prioritized. This priority state information indicates “1” when priority is given to viscosity, “2” when priority is given to stiffness, “3” when priority is given to burning, and “0” when neither value is given priority. In the object information ID number, as shown in FIG. 44B, the name of the object information is recorded for each object information ID. The viscosity is expressed in five stages from “1” to “5”. “1” indicates that the viscosity value is small, for example, a state where there is no stickiness of the material indicated by the object information. When the viscosity value is 5, it indicates a state in which the stickiness is considerably strong. The rigidity is shown in five stages from “1” to “5”. When the stiffness value is “1”, the material indicated by the object information is soft, and when it is “5”, it is hard. Burning is shown in five stages from “1” to “5”, “1” indicates that there is little burning, and “5” indicates that burning is large. The progress information is information indicating whether or not the current operation is a process, and “1” indicates that the current operation is being performed, and “0” indicates that the operation is not being performed.

44A shows a series of operation IDs in the operation database 17, and shows the operation of the robot arm 5 in each process.

As shown in FIG. 45, the work procedure information database 18 collects home work procedure information such as cooking recipe information as an example of work from an information database 98 in an external web server via the Internet 99. Alternatively, the manufacturer of the robot arm 5 may prepare in advance in the control device of the robot arm 5 when the robot arm 5 is shipped.

FIG. 35 is a specific example of the operation database 17 of the second embodiment. Since the description about each information is the same as that of 1st Embodiment, description is abbreviate | omitted. However, since the information of “suction force” is not used for the operation of stirring the pot 3, all “0” is stored for “suction force”, and “suction” of “flag” and “correction parameter flag” is stored. “0” is stored for the “force” bits (12th to 17th bits).

35, operation IDs “1” to “8” are operations for stirring the whole of the pot 3 while rubbing the pot bottom, as shown in FIG. 36A.

Information on the position, posture and time of the robot arm 5 in the motion database 17 is obtained by, for example, directly holding the robot arm 5 with a human hand 16 as shown in FIG. 5, the control unit 22 obtains information on the hand position and posture of the robot arm 5 at certain time intervals (for example, every 0.2 msec), and together with the time information, the operation database 17 stores the information on the operation database 17. Remember. Note that information on position, orientation, and time may be generated in advance by the manufacturer at the time of product shipment and stored in the operation database 17.

The object state determination unit 19 determines the state of the object according to information regarding the state of the work procedure information database 18 with respect to the process ID in which the robot arm 5 is operating in the work procedure information database 18 (for example, The state of the ingredients in the pan 3 as an example of the object is determined in terms of the viscous state, the rigid state, and the burned state). Specifically, the state ID with the process ID “1” in FIG. 44A is “1”, and the priority state with the state ID “1” in FIG. 44C is “0”. Assuming that the state determination unit 19 does not make a determination, the object state determination unit 19 outputs “0”. When the process ID in FIG. 44A is “2”, the state ID is “2”, and the priority state is “3” from FIG. 44C. “Scorching” when the state ID is “2” has a small value of “1” (less than “3”, which is an example of a scoring threshold value). The object state determination unit 19 outputs “0”. When the state ID in FIG. 44C is “3”, the priority state is “3”, and the burnt state is prioritized. “Burn” when the state ID is “3” is “3”. Therefore, when the burn value is “3” or more, which is an example of the threshold value for burn, the burned state and the object The state determination unit 19 makes a determination, and the object state determination unit 19 outputs “1”. Since the priority states in the case where the state ID in FIG. 44C is “5” and the state ID is “9” respectively indicate “1”, the viscous state is prioritized. When “viscosity” is 2 or less (less than “3”, which is an example of a threshold value for viscosity), the object state determination unit 19 determines that the viscosity is small, and the object state determination unit 19 sets “0”. Output. When the “viscosity” is 3 or more (when it is “3” or more, which is an example of a threshold value for viscosity), the object state determination unit 19 determines that the viscosity is large, and the object state determination unit 19 sets “1”. Output. Since the priority states in the case of the state ID “6” and the state “7” in FIG. 44C are “2”, the rigidity state is given priority. When the value of “rigidity” is less than “3”, which is an example of the threshold value for rigidity, that is, 2 or less, the object state determination unit 19 determines that the rigidity is small, that is, the soft state, and the target The object state determination unit 19 outputs “1”. When the value of “rigidity” is equal to or greater than “3”, which is an example of a stiffness threshold value, the object state determination unit 19 determines that the object is in a hard state, and the object state determination unit 19 sets “0”. Output. As described above, the object state determination unit 19 compares the numerical value representing each state of the object with the threshold value related to each state, and changes the output value at or above the threshold value. Judging the state.

43 selects an optimum work from the work list in the work procedure information database 18 (for example, work displays such as “stirring” and “wiping cleaning” displayed in the center of the cleaning switch 26c in FIG. 26). When the person 16A selects via the data input IF 26, “1” is set to the progress information of the currently operating process ID among the selected works, and is stored in the work procedure information database 18. For the process, “0” is set and stored in the work procedure information database 18. That is, the operation selection unit 29 measures the time with the built-in timer of the control device 1000 from the time of selecting the work, and based on the elapsed time and information on the time during which each operation of the work procedure information database 18 acts. Of the selected work, “1” is set in the progress information of the currently operating process ID and stored in the work procedure information database 18, and “0” is set for the other processes and the work procedure information. Store in database 18. The operation selection unit 29 refers to the operation ID of the selected process ID from the operation database 17 and sets “1” in the progress information of the currently operating operation ID.

As in the first embodiment, the correction operation type determination unit 23 is a correction type that allows the operation correction unit 20 to correct an operation by applying a force to the robot arm 5 with a hand 16 of the person. To decide. Hereinafter, there are five types of correction.

The first correction type is “movement of the position of the work surface”. Specifically, as shown in FIGS. 38A and 38B (viewed from the top of FIG. 38A), when the stirring operation is performed while rubbing the bottom of the pan 3, using the operation information that operates, FIG. As shown in FIG. 38, when a force is applied to the robot arm 5 from the side with the human hand 16, the position of the robot arm 5 in the horizontal direction with respect to the work surface of the robot arm 5 (the bottom of the pan 3) as shown in FIG. The ladle 4 held by the hand 30 of the robot arm 5 can be moved in parallel with the pan bottom surface of the pan 3.

The second type of correction is the “force applied” when the pot bottom is stirred. This is effective when the force bit is “1” in the operation flag (the flag indicating validity) of the operation currently being performed (the progress information of the operation database 17 is “1”). As shown in FIG. 36A, when a force is applied downward from above to the robot arm 5 with a human hand 16 as shown in FIG. On the contrary, when a force is applied to the robot arm 5 upward from below with the human hand 16, the mixing operation of the robot arm 5 can be corrected to weaken the force application.

The third type of correction is the “speed” of stirring (movement) of the hand of the robot arm 5 (ie, the ladle 4). As shown in FIG. 40A, when the robot arm 5 is being stirred, if a force is applied to the robot arm 5 with a human hand 16 in a direction opposite to the advancing direction of the robot arm 5 as shown in FIG. As shown in FIG. 40C, the moving speed during the stirring operation can be reduced. On the contrary, when the robot arm 5 is in operation and the human hand 16 applies force to the robot arm 5 with the human hand 16 in accordance with the moving direction of the robot arm 5, the movement during the mixing operation is performed by the motion correction unit 20. Speed can be accelerated.

The fourth type of correction is “change of direction (posture)”. As shown in FIG. 39A, while the robot arm 5 is performing the stirring operation using the motion information to be operated, a force is applied to change the orientation of the hand 30 of the robot arm 5 with the human hand 16 as shown in FIG. 39B. Is applied, the motion correcting unit 20 can change the posture of the hand 30 of the robot arm 5 as shown in FIG. 39C and change the motion to rub the inner side surface of the pan 3 with the ladle 4. This can be realized by changing the posture (φ, θ, ψ) of the hand (hand 30) of the robot arm 5.

The fifth type of correction is “movement in the vertical direction of the work surface”. As shown in FIG. 41A, during the agitation work with the robot arm 5, as shown in FIG. 41B, an upward force is applied to the robot arm 5 with a human hand 16 to place the robot arm 5 on the convex portion 7a of the work table 7. When the robot arm 5 is moved upward toward the pan 3a, the operation correction unit 20 can perform the stirring work with the ladle 4 in the pan 3a as shown in FIG. 41C.

The correction operation type determination unit 23 determines one type of correction among the five types of correction. Specifically, one of the five types of correction is selected by the data input IF 26 such as a button, or is detected by the force detection unit 53 and acquired by the information acquisition unit 100. The force applied to the robot arm 5 by the human hand 16, the force applied to the robot arm 5 stored in the motion database 17 and acquired by the information acquisition unit 100, and the state of the object detected by the object state determination unit 19 Then, the correction operation type determination unit 23 estimates the type based on the relationship information with the correction type (for example, the relationship information between the direction in which the force is applied, the magnitude, and the correction type).

If the power button 26a of the household robot 1 is set to “ON” and the robot arm 5 is not gripped by the human hand 16 and no force is applied to the robot arm 5, the robot arm 5 does not move. When a force is applied to the robot arm 5 by the human hand 16, the robot arm 5 is moved in the direction in which the robot arm 5 is moved in the impedance control mode (a mode in which the force of the human hand 16 is detected and moved by impedance control). Can be made. In this case, the force detection unit 53 of the control unit 22 detects the force acting on the robot arm 5, and the information on the force detected by the force detection unit 53 is determined via the information acquisition unit 100. The data is input to the unit 23 (step S71).

Next, in step S72, the all components obtained in is detected by the force detection unit 53 and the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is (specifically, (f _dx of ID "1" in FIG. _{_{33, f dy, f dz,}} f dφ, f dθ, f dψ)) is the threshold correction or not less whether the operation type determining unit 23 Judge with. All components were detected by the force detection unit 53 and acquired by the information acquisition section 100 a force _{_{_{(f x, f y, f}}} z, f φ, f θ, 6 components of f _[psi) is below the certain threshold If the correction action type determination unit 23 determines that there is, the robot arm 5 does not move and is not corrected (step S88), and the correction type estimation process of the correction action type estimation method ends. The control mode in that case is an impedance control mode.

At step S72, the one of component _(f _x of the acquired detected by the force detection unit 53 and the information acquisition unit 100 _{force, f y, f z, f} φ, f θ, of the six components of the f _[psi ) _{Exceeds a} certain threshold (specifically, (f _dx , f _dy , f _dz , f _dφ , f _dθ , f _dψ ) of ID “1” in FIG. 33). If the type determining unit 23 determines, the process proceeds to step S73.

In step S73, the corrected operation type determination unit 23 further determines whether or not the current household robot 1 is operating in the operation database 17 based on the information acquired through the information acquisition unit 100. Specifically, the correction operation type determination unit 23 determines that no operation is selected by the operation selection unit 29 and the progress information is “0” for all the operation IDs in the operation database 17. In the case of determination (a state in which work is not started), the correction operation type determination unit 23 determines that the operation database 17 is not operating, and does not perform correction (step S77). The correction type estimation process ends. When the mixing operation is selected by the operation selection unit 29 and the mixing operation is started, and when the correction operation type determination unit 23 determines that the progress information is “1”, the operation database 17 The correction operation type determination unit 23 determines that the operation is in progress, and proceeds to step S74.

In step S74, when the robot arm 5 is gripped by the human hand 16 and a force is applied in a direction in which the operation of the robot arm 5 is desired to be corrected, the force applied to the robot arm 5 is detected by the force detection unit 53, and the force is detected. Displacement of the forces (f _x , f _y , f _z , f _φ , f _θ , f _ψ ) from the human hand 16 detected by the detection unit 53 and acquired through the information acquisition unit 100 for a certain period of time. The amount is measured by the correction action type determination unit 23, and the correction action type indicates which displacement amount of the position component (f _x , f _y , f _z ) or the posture component (f _φ , f _θ , f _ψ ) is larger. Measurement is performed by the determination unit 23. Specifically, as shown in FIG. 15, each time series force (f _x , f _y , f _z , f _φ , f _θ , f _ψ ) is measured by the correction operation type determination unit 23, and is constant. The correction operation type determination unit 23 measures how much the force is displaced in time (for example, time 1), and the correction operation type determination unit 23 measures the component having the largest displacement. In this example, the displacement of f _phi largest, as judged by the correction operation type determining unit 23 orientation component is force is applied from the position component, the process proceeds to step S79.

When the correction operation type determination unit 23 determines in step S74 that the displacement amount of the posture is larger than the displacement amount of the position, the correction operation type determination is made that the correction type is “direction (posture) change”. Then, the correction type estimation process is completed (step S79). The control mode at that time is the same control mode (force hybrid impedance control mode) as before the correction type is determined.

On the other hand, when the correction operation type determination unit 23 determines in step S74 that the displacement amount of the position is equal to or greater than the displacement amount of the posture, the force component in the direction perpendicular to the work surface (the bottom surface of the pan 3) Correction operation (for example, f _z when stirring along the bottom of the pan 3 placed horizontally on the ground) is greater than a certain threshold (specifically, f _dz of ID “1” in FIG. 33) The type determination unit 23 determines (step S75).

In step S75, when it is determined by the correction operation type determination unit 23 that the force component in the direction perpendicular to the work surface (the bottom surface of the pan 3) is smaller than the certain threshold value, the work surface (the bottom surface of the pan 3) Force component in a horizontal direction (for example, f _x and / or f _y when stirring along the bottom of the pan 3 horizontally disposed on the workbench 7) has a certain threshold (specifically, The correction operation type determination unit 23 determines whether or not it is greater than or _{equal to} f _dx , f _dy of ID “1” in FIG. 33 (step S80).

In step S80, (specifically, f _x of ID "1" in FIG. _33, f _y) working surface thresholds horizontal force component (pot bottom of the pot 3) is said to be less than the correction When the operation type determination unit 23 determines, no correction (no type) is determined, and the correction type estimation process ends (step S81). If there is no correction, stop the correction and work.

In step S80, if the correction operation type determination unit 23 determines that the force component in the direction horizontal to the work surface (the bottom surface of the pan 3) is equal to or greater than the certain threshold value, the process proceeds to step S91.

In step S91, the object state determination unit 19 determines the state of the object based on the information in the work procedure information database 18. When “1” is output from the object state determination unit 19, that is, when the object state determination unit 19 determines that the object is burnt, easily crushed, or has a strong stickiness, “Speed correction” is determined by the correction operation type determination unit 23 (step S85), and the correction type estimation process ends. If the object state determination unit 19 outputs “1” in step S91, that is, if it is determined that the object is scorched, crushed easily, or has a strong stickiness, the speed is determined in step S85. Correction can be made while adjusting the adjustment. For example, if the object is burnt, the speed can be increased. If the object is easily crushed, the speed can be decreased. If the object is strong, the object can be mixed at an increased speed.

In step S91, when the object state determination unit 19 determines the state of the object based on the information in the work procedure information database 18, and outputs “0” from the object state determination unit 19, that is, burns. If the object state determination unit 19 determines that the object state is not hard, or the stickiness is small, the process proceeds to step S83.

In step S83, the horizontal movement amount of the work surface (the bottom surface of the pan 3) calculated by the correction operation type determination unit 23 is set to a certain threshold value (specifically, g _x of ID “2” in FIG. 33). , G _y ) or more, when the correction operation type determination unit 23 determines that the correction type is “movement of the work surface position”, the correction operation type determination unit 23 determines the correction type estimation process. Is finished (step S84). In addition, when the movement amount in the horizontal direction of the work surface (the bottom surface of the pan 3) is calculated by the correction operation type determination unit 23, specifically, the control unit 22 through the control parameter management unit 21 or the information acquisition unit 100. Then, the hand position of the robot arm 5 before the human operation and the hand position during the operation are input to the correction operation type determination unit 23, and a correction operation is performed using a value obtained by subtracting the hand position before the operation from the hand position during the operation. It can be calculated by the type determining unit 23. Further, when the movement amount in the direction perpendicular to the work surface (the pan bottom surface of the pan 3) is calculated by the correction operation type determination unit 23, specifically, the control parameter management unit 21 or the information acquisition unit 100 is controlled from the control unit 22. Then, the z component of the hand position of the robot arm 5 before the human operation and the z component of the hand position during the operation are input to the correction operation type determination unit 23, and the hand position before the operation is determined from the z component of the hand position during the operation. A value obtained by subtracting the z component of can be calculated by the correction operation type determination unit 23 as a movement amount.

In step S83, when the correction operation type determination unit 23 determines that the horizontal movement amount of the work surface (the bottom surface of the pan 3) is less than the certain threshold value, the correction surface is horizontal as the correction type. The type of “speed” in the correct direction is determined, and the correction type estimation process is terminated (step S85).

In step S75, if the correction operation type determination unit 23 determines that the force perpendicular to the work surface (the bottom surface of the pan 3) is equal to or greater than the certain threshold value, the process proceeds to step S90.

In step S90, the object state determination unit 19 determines the state of the object based on the information in the work procedure information database 18. When “1” is output from the object state determination unit 19, that is, when the object state determination unit 19 determines that the object is burnt, easily crushed, or has a strong stickiness, The “force correction” is determined by the object state determination unit 19 (step S86), and the correction type estimation process ends. If the object state is determined by the object state determination unit 19 in step S90 and “0” is output, that is, if it is determined that the object is not burnt, is hard, or has a small stickiness, the process proceeds to step S82. move on.

In step S82, if the correction operation type determination unit 23 determines that the vertical movement amount of the work surface (the bottom surface of the pan 3) is equal to or less than the certain threshold value, the process proceeds to step S86.

In step S90, when the state of the object is determined by the object state determination unit 19 and "0" is output, that is, the object state is determined to be burnt, easily crushed, or strong sticky. When determining by the unit 19, the correction operation type determination unit further determines whether or not the vertical movement amount of the work surface (the bottom surface of the pan 3) calculated by the correction operation type determination unit 23 is greater than a certain threshold value. The determination is made at step 23 (step S82).

In step S82, when the correction operation type determining unit 23 determines that the vertical movement amount of the work surface (the bottom surface of the pan 3) is larger than the certain threshold value, the correction type is “work surface vertical direction”. The movement type "is determined by the correction operation type determination unit 23, and the correction type estimation process is terminated (step S87).
By this step S90, when the object is scorched, is in a state of being easily crushed, or has a strong stickiness, correction can be performed while adjusting the force. For example, if it is burnt, the power can be increased, if it is crushed easily, the power can be weakened, and if it is sticky, the power can be increased to mix well.

The correction operation type determination unit 23 determines one type of the above five types, but can also determine two types of correction at the same time.

The correction operation type determination unit 23 determines the correction type according to the number of outputs set by the correction operation type determination method setting unit 27. Specifically, when the number of outputs is 1, the correction type is determined by the algorithm of the correction type estimation method in FIG. As a result, when the person who operates the household robot 1 is not familiar with the operation, setting the number of outputs to 1 makes it impossible to perform two types of correction at the same time, thus simplifying the operation. On the other hand, if you are used to the operation and want to correct two types at the same time, the correction can be performed efficiently by setting the number of outputs to a value of “2”.

The correction operation type determination unit 23 described above outputs one type, but it may output and correct two types as in the first embodiment.

Similar to the first embodiment, the motion correction unit 20 applies the force to the robot arm 5 with the human hand 16 during the motion based on the position, posture, and time of the motion database 17, and thus the motion of the motion database 17. This function corrects information.

Hereinafter, the function of the motion correction unit 20 will be described.

Next, with the human hand 16, the operation selection unit 29 selects a desired work from the work list in the work procedure information database 18 and gives an instruction to start the operation. The motion correction unit 20 is based on the motion information of the motion ID indicated by the process ID in the work procedure information database 18 (specifically, the position of the rail movable portion 8b and the position, posture, and time of the robot arm 5). The control mode of 8b and the robot arm 5 is set. In this example, since the work ID “3” in FIG. 35 is selected, the flag bit is “1” in the flag (the flag indicating validity) for the operation ID “1” in the operation database 17. A hybrid impedance control mode (a mode in which the robot arm 5 operates according to the force applied to the robot arm 5 from a person or the like while operating in the position control mode) for each position and posture of the robot arm 5 ) Is set by the operation correction unit 20, and a command is issued to the control parameter management unit 21. When a command is issued from the operation correction unit 20 to the control parameter management unit 21, a stirring operation is started as shown in FIG. 38A.

In the case of the force hybrid impedance control mode, among the flags for the action IDs in the action database 17 (flags indicating validity), for each position and posture of the robot arm 5 in which the flag bit is “1”. A hybrid impedance control mode (a mode in which the robot arm 5 is actuated according to the force applied to the robot arm 5 by a person or the like while operating in the position control mode) is set by the motion correction unit 20. The force control mode is set by the operation correction unit 20 for the component in which the bit of the flag (flag indicating validity) is “1”. Of the six components of position and orientation, the component for which neither the hybrid impedance control mode nor the force control mode is set has the impedance control mode set by the operation correction unit 20.

For example, the work ID “3” in FIG. 35 shows the work of stirring while rubbing the bottom of the pot, and the flag when the operation ID is “1” is “1” only in the first, second, and eighth bits. The hybrid impedance control mode is set by the motion correction unit 20 for the x-axis component and the y-axis component, and the force control mode is set by the motion correction unit 20 for the z-axis component. Thus, the impedance control mode is set by the operation correction unit 20.

The control parameter management unit 21 receives a command from the operation correction unit 20. That is, when a command is issued from the operation correction unit 20 to the control parameter management unit 21 to perform the mixing operation in the force hybrid impedance control mode, the mixing operation is started as shown in FIG. 38A.

Next, as shown in FIG. 38D, an example will be described in which the robot arm 5 is desired to be moved in the horizontal direction for a while.

As shown in FIG. 38C, the robot arm 5 is directly gripped by a human hand 16 and is parallel to the work surface (the bottom surface of the pan 3) so as to move parallel to the work surface (the bottom surface of the pan 3). Apply force to the robot arm 5.

The correction type estimation shown in the flowchart of FIG. 42 based on the force applied by the human hand 16 to the robot arm 5 and the information stored in the action database 17 acquired by the correction action type determination unit 23 by the information acquisition unit 100. The type of correction is estimated and determined by processing. While operating in the process ID “2”, when the human arm 16 moves the robot arm 5 by more than a certain threshold by applying a force to the robot arm 5 in a direction parallel to the work surface (the bottom surface of the pan 3). In step S90, since the priority information of the state “2” in the case of the process ID “2” is “3”, priority is given to scorching. Since the burning is “1” which is 2 or less (less than “3” which is an example of a burning threshold), the object state determination unit 19 outputs “0”. If “0” in step S90, the process proceeds to step S82. Since the movement exceeds the threshold value in step S82, “movement in the vertical direction of the work surface” is determined. When a force is applied to the robot arm 5 in the direction parallel to the work surface (the bottom surface of the pan 3) with the human hand 16 during operation in the process ID “3”, the state is changed to “3” in step S90. ”, Since the priority information is“ 3 ”, priority is given to scorching. Since the scoring is “3” (“3” or more, which is an example of a scoring threshold value), the object state determination unit 19 determines “1”, and the object is scoring. "Force correction". As a result, when it is in a burned state, it can be stirred with a stronger correction of power, so that it is possible to suppress scorching.

Also, when operating with the process ID “6”, the state of the process ID [6] is “6” and the priority information is “3”, that is, the rigidity is given priority. The rigidity of the state ID “6” is “3”, which is equal to or greater than “3”, which is an example of a stiffness threshold value, and therefore the object state determination unit 19 outputs “0”. Since “0” in step S90, “force correction” or “movement in the vertical direction of the work surface” is determined according to the amount of movement. When operating with the process ID “7”, the priority information of the state “7” gives priority to “3”, that is, rigidity. The stiffness value is “1”, which is easily crushed, and is less than “3”, which is an example of the stiffness threshold value. Therefore, the object state determination unit 19 outputs “1”. Next, in step S86, the correction operation type determination unit 23 determines that the correction type is “force correction”. As a result, “potatoes” that tend to be crushed can be mixed so that they will not be crushed by correcting the force.

In addition, a case will be described in which a force is applied to the work surface in the horizontal direction to move it beyond the threshold during operation with the process ID “5” or process ID “9”. Here, the case where curry or stew is made will be described as a more specific example of the cooking operation. The process ID “5” and the process ID “9” are processes before and after adding the curry roux (or stew roux) and water, respectively. In each state, priority information “1”, that is, viscosity is given priority. In the process ID “5”, the viscosity value is “1”, which is less than “3”, which is an example of a threshold value for viscosity. Therefore, the object state determination unit 19 outputs “0”. In step S83, since the movement amount is equal to or larger than the threshold value, “movement of the position of the work surface” is determined in step S84. In the process ID “9”, the viscosity value is “4”, which is equal to or greater than “3”, which is an example of a threshold value for viscosity, so the object state determination unit 19 outputs “1”. Next, in step S85, “speed correction” is determined.

As described above, when the curry roux (or stew roux) is put into the pan and the stickiness of the liquid in the pan is strong, the mixing speed can be corrected quickly to make it possible to mix well. .

35, when the work ID is “3” and the action ID is “1”, the x-axis component and the y-axis component are moved by the force hybrid impedance control mode while moving the robot arm 5 in the position control mode. The force applied to the robot arm 5 by the human hand 16 in the impedance control mode is detected by the force detection unit 53, and the robot arm 5 is moved in the x-axis direction and in the direction in which the force is applied to the robot arm 5 by the human hand 16. By moving in the y-axis direction, the position can be corrected as shown in FIG. 38D.

In this example, since it is desired to correct the operation only in the x-axis direction and the y-axis direction, the correction operation type determination unit 23 performs the correction operation type determination unit 23 at the timing when the correction type is determined. By setting the 0th and 1st bits of the correction parameter flag to “1” and setting the other bits to “0” and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, It can be set so that movements other than the direction of the x-axis and the direction of the y-axis cannot be performed. Furthermore, by changing the mechanical impedance setting value in the impedance control mode by the correction operation type determination unit 23 and issuing a command from the correction operation type determination unit 23 to the control parameter management unit 21, the x-axis direction and y of the robot arm 5 The rigidity in the axial direction is made lower than the rigidity in the other directions so that the robot arm 5 can be easily moved in the x-axis direction and the y-axis direction by the human hand 16, and the rigidity in directions other than the x-axis direction and the y-axis direction can be increased. The robot arm 5 can be made difficult to move in a direction other than the x-axis direction and the y-axis direction by the human hand 16. Thereby, when it is desired to correct only the x-axis component and the y-axis component of the robot arm 5, it is possible to prevent the z-axis component of the robot arm 5 from being erroneously corrected. Further, during the correction of the robot arm 5 in the x-axis direction and the y-axis direction, the force applied to the z-axis component by the correction operation type determination unit 23 is weaker or smaller than the operation before the correction (specifically, about half). You can also Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. Accordingly, it is possible to prevent the pot 3 from being damaged by applying a force to the robot arm 5 during the correction by moving the robot arm 5 in the x-axis direction and the y-axis direction.

As described above, the robot arm 5 is gripped by a human hand 16 and a force is applied in a direction parallel to the work surface (the bottom surface of the pan 3) to cause the robot arm 5 to move in the x-axis direction by Δx and Δy. When moved in the y-axis direction, the value of Δx and the value of Δy are transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 subtracts Δx from all the x coordinate values of the motion information of the selected work ID, and further subtracts Δy from all the y coordinate values to correct the motion information corrected by the motion correction unit 20. To the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates for Δx and Δy. This corrects the operation as shown in FIG. 38D. Next, the operation information reduced by Δx and Δy is stored in the operation database 17 by the operation storage unit 15.

Next, as shown in FIG. 41B, for example, when the mixing work of the pot 3 is performed with another pot 3a, the robot arm 5 is directly gripped by a human hand 16 and the work surface (the pot 3 A force is applied to the robot arm 5 perpendicular to the work surface (the bottom of the pan 3) so as to move in a direction perpendicular to the bottom of the pan).

Based on the force applied by the human hand 16 to the robot arm 5 and the information in the motion database 17 respectively acquired by the information acquisition unit 100 by the correction operation type determination unit 23, the correction type estimation process shown in the flowchart of FIG. The type of correction is estimated and determined. Here, since the robot arm 5 is moved by the human hand 16 in a direction perpendicular to the work surface (the bottom surface of the pan 3) with the human hand 16 to move the robot arm 5 by more than the predetermined threshold value, in step S87, the correction is performed. The correction operation type determination unit 23 determines that the type is “movement of the work surface in the vertical direction”.

While the robot arm 5 is moved in the position control mode by the force hybrid impedance control mode, the force of the human hand 16 is detected by the force detection unit 53 by the impedance control mode and the force is applied to the robot arm 5 by the human hand 16. The agitation position can be corrected as shown in FIG. 41C by moving the robot arm 5 in the z-axis direction in the applied direction.

Further, when correcting the movement of the robot arm 5 in the z-axis direction, the correction action type determination unit 23 causes the force applied to the work surface of the z-axis component to be weaker or smaller than that before the correction (specifically, It can also be about half). Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. Thereby, it is possible to prevent the pot 3 from being damaged by applying a force to the robot arm 5 while moving the robot arm 5 in the z-axis direction.

As described above, the robot arm 5 is grasped by the human hand 16 and the robot arm 5 is moved in the z-axis direction by Δz by applying a force in a direction perpendicular to the work surface (the bottom surface of the pan 3). In this case, the value of Δz is transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 transmits motion information corrected by subtracting Δz from all z coordinate values of the motion information of the selected work ID from the motion correction unit 20 to the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates corresponding to Δz. This corrects the operation as shown in FIG. 41C. Next, the motion information reduced by Δz is stored in the motion database 17 by the motion storage unit 15.

For example, when the robot arm 5 is changed in posture as shown in FIG. 39C, the robot arm 5 is directly held by a human hand 16 and the robot arm 5 is moved in the direction to be changed as shown in FIG. 39B. Move.

Based on the force applied by the human hand 16 to the robot arm 5 and the information in the motion database 17 respectively acquired by the information acquisition unit 100 by the correction operation type determination unit 23, the correction type estimation process shown in the flowchart of FIG. The type of correction is estimated and determined. Here, since force is applied to move the robot arm 5 in the direction in which it is desired to be changed by the human hand 16, in step S79, if the type of correction is “change in direction (posture)”, the correction operation is performed. The type is determined by the type determining unit 23.

While the robot arm 5 is moved in the position control mode by the force hybrid impedance control mode, the force applied to the robot arm 5 by the human hand 16 is detected by the force detection unit 53 by the impedance control mode. The direction can be corrected as shown in FIG. 39C by rotating the robot arm 5 in the φ axis direction in the direction in which a force is applied to the robot arm 5.

In this example, since it is desired to correct the operation only in the direction of the φ axis, the correction operation type determination unit 23 sets the correction parameter flag in FIG. 6 at the timing when the correction type is determined by the correction operation type determination unit 23. The third bit is set to “1” and the other bits are set to “0”, and a command is issued from the correction operation type determination unit 23 to the control parameter management unit 21. As a result, the correction operation type determination unit 23 can set the movement not to move in the direction other than the φ-axis direction. Further, the mechanical impedance set value in the impedance control mode is changed by the correction operation type determination unit 23, and a command is issued from the correction operation type determination unit 23 to the control parameter management unit 21, so that the rigidity in the φ-axis direction is changed to another direction. The rigidity of the robot arm 5 is made easier to move in the φ-axis direction by the human hand 16 and the rigidity other than the φ-axis direction is increased, and the robot arm 5 is moved in the direction other than the φ-axis direction by the human hand 16. It can be made difficult to move.

Further, during the correction of the robot arm 5 in the φ-axis direction, the force applied to the z-axis component work surface (the bottom surface of the pan 3) by the correction operation type determination unit 23 is weaker or smaller than that during the operation before the correction ( Specifically, it can be about half). Alternatively, a command can be issued from the correction operation type determination unit 23 to the control parameter management unit 21 so as to stop the force control. Specifically, the 6th to 17th bits of the flag of the operation database 17 are set to “0” by the correction operation type determination unit 23. Thereby, it is possible to prevent the bottom of the pan 3 from being damaged by applying force to the robot arm 5 during the movement in the φ axis direction.

As described above, the robot arm 5 is grasped by the human hand 16 and the robot arm 5 is rotated in the φ axis direction by Δφ by applying a force in a direction perpendicular to the work surface (the bottom of the pan 3). In this case, the value of Δφ is transmitted to the operation correction unit 20 via the control unit 22 and the control parameter management unit 21.

The motion correction unit 20 transmits the motion information corrected by subtracting Δφ from all the φ coordinate values of the motion information of the selected work ID from the motion correction unit 20 to the control parameter management unit 21. The control parameter management unit 21 instructs the control unit 22 to operate the robot arm 5 with the corrected coordinates corresponding to Δφ. This corrects the operation as shown in FIG. 39C. Next, the operation information reduced by Δφ is stored in the operation database 17 in the operation storage unit 15.

As described above, the motion correction unit 20 applies force to the robot arm 5 with the human hand 16 while operating in the hybrid impedance control mode or the force hybrid impedance control mode depending on the position, posture, and time of the motion database 17. By applying, the generated position can be corrected for each direction.

Next, as shown in FIG. 36C, when the force on the work surface (the bottom surface of the pan 3) at the time of the agitation work is changed, the robot arm 5 is directly gripped by the human hand 16 and the work surface (the pan 3) A force is applied to the robot arm 5 in a direction perpendicular to the bottom of the pan).

Based on the force applied by the human hand 16 to the robot arm 5 and the information in the motion database 17 respectively acquired by the information acquisition unit 100 by the correction operation type determination unit 23, the correction type estimation process shown in the flowchart of FIG. The type of correction is estimated and determined. Here, since a force is applied to the robot arm 5 in the direction perpendicular to the work surface (the bottom surface of the pan 3) with the human hand 16 and the robot arm 5 is not moved beyond the certain threshold value, the type of correction is determined in step S86. The correction action type determination unit 23 determines that the type is “force correction”.

At the timing when the correction operation type determination unit 23 determines that the correction type is “force correction”, the correction operation type determination unit 23 controls the control parameter management unit to operate in the high-rigidity position control mode from the force hybrid impedance control mode. A command is issued to 21. In the high-rigidity position control mode, the high-rigidity at the time of position control can be set in the high-rigidity position control mode by the correction-operation-type determining unit 23 at the time of command from the correction operation type determining unit 23 to the control parameter management unit 21. Since the operation flag of operation ID “3” and operation ID “1” in the operation database 17 is set to “1” in the 0, 1, and 8 bits, the z-axis direction operates in the force control mode. Since the other directions operate in the hybrid impedance control mode, the control parameter management unit 23 controls the control parameters so that only the z-axis direction operates in the high-rigidity position control mode and the other directions operate in the hybrid impedance control mode. A command is issued to the unit 21.

Next, as shown in FIG. 36B, when the robot arm 5 is performing the mixing operation, and the robot arm 5 is directly held by the human hand 16 and it is desired to change the mixing force to be stronger. Applies a downward force to the robot arm 5 toward the work surface (the bottom surface of the pan 3) with a human hand 16. In the high-rigidity position control mode, the hybrid impedance control mode is a mode in which the position control mode set for each direction is further increased in rigidity, and the gain in the position error compensation unit 56 is increased (specifically, When the force is applied to the robot arm 5 with the human hand 16, the robot arm 5 can be prevented from moving easily. Thus, the force applied to the robot arm 5 can be detected by the force detection unit 53. The force detected by the force detection unit 53 of the control unit 22 is notified to the operation correction unit 20. By storing the force notified to the motion correction unit 20 in the motion database 17 in the motion storage unit 15, the motion can be corrected so as to stir strongly.

When no force is applied to the robot arm 5 with the human hand 16, all the components of the force are equal to or less than the threshold values in step S72 of FIG. 42, so that the correction operation type determination unit 23 sets “correction” as the correction type. "None" (step S88 in FIG. 42). The motion correction unit 20 receives the information of “no correction” and issues a command from the correction operation type determination unit 23 to the control parameter management unit 21 so as to control from the highly rigid position control mode to the hybrid impedance control mode. Thus, the mixing operation is performed in the corrected operation database 17.

As described above, the motion correction unit 20 works with the corrected force by applying the force of the human hand 16 in the state of operating in the hybrid impedance control mode based on the force information of the motion database 17. It becomes possible to correct.

Next, as shown in FIG. 40B, when the mixing speed is changed, the robot arm 5 is directly held by the human hand 16 and accelerated. When it is desired to decelerate, the human arm 16 applies force to the robot arm 5 in the direction opposite to the traveling direction. At this time, the speed of the hand position of the robot arm 5 may be changed, but the force is applied to the robot arm 5 with the human hand 16 so as not to move the position beyond the certain threshold.

The correction type estimation process shown in the flowchart of FIG. 42 is performed based on the force applied to the robot arm 5 with the human hand 16 and the information in the action database 17 respectively acquired by the correction action type determination part 23 by the information acquisition part 100. The type of correction is estimated and determined. Here, since a force is applied to the robot arm 5 in the direction horizontal to the work surface (the bottom surface of the pan 3) with the human hand 16 and the robot arm 5 is not moved beyond the certain threshold, step S85 in FIG. As the type of correction, the correction operation type determination unit 23 determines that the type is “speed” in the direction horizontal to the work surface (the bottom of the pan 3).

While the robot arm 5 is moved in the position control mode by the hybrid impedance control mode, the force applied to the robot arm 5 by the human hand 16 is detected by the force detection unit 53 by the impedance control mode, and the robot is moved by the human hand 16. The robot arm 5 is moved in the x-axis direction and the y-axis direction in the direction in which a force is applied to the arm 5. From the position of the robot arm 5 shown in the operation database 17, for example, work ID operation ID as _{_{_{(x 1, y 2, z}}} 1), the position of the robot arm 5 of the next operation _{_{ID (x 2, y 2,}} z ₂ ) If the time taken to move to t) is t ₁ , when the speed of the robot arm 5 is changed by the force of the human hand 16 (see FIG. 40B), that is, the position (x ₁ , y ₂ , z ₁ ) to the position (x ₂ , y ₂ , z ₂ ), when the time taken to move from t ₁ to t ₂ is changed, the value of time t ₂ is controlled by the control unit 22 and control parameter management. It is transmitted to the operation correction unit 20 via the unit 21. The motion correction unit 20 changes the motion information of the selected work ID from the time t _{1 to} the time t ₂ and transmits the motion information from the motion correction unit 20 to the control parameter management unit 21. In the control parameter managing unit 21 instructs the control parameter managing unit 21 to operate in t ₂ is corrected time to the control unit 22. This corrects the operation as shown in FIG. 40C. Next, the time t ₂ is stored in the operation database 17 by the operation storage unit 15.

It should be noted that, by appropriately combining any of the various embodiments or modifications, the effects possessed by them can be produced.

The present invention relates to a robot arm control device and control method, a house robot, and a robot arm control for controlling the operation of a robot arm of a house robot when a person and a robot work together such as a home robot. It is useful as an integrated electronic circuit for program and robot arm control.

Although the present invention has been fully described in connection with preferred embodiments with reference to the accompanying drawings, various variations or modifications will be apparent to those skilled in the art. Such changes and modifications are to be understood as being included therein unless they depart from the scope of the invention as defined by the appended claims.

Claims

A control device for a robot arm that controls the operation of the robot arm and performs housework work while acting on the object of housework work by the robot arm in the home,
Force detecting means for detecting the force of a person acting on the robot arm;
An information acquisition unit for acquiring information on the operation including the position of the robot arm in the housework and information on the human force detected by the force detection unit;
Object state determination means for determining the state of the object;
The motion from the information regarding the motion including the position of the robot arm in the housework work acquired by the information acquisition unit, the information regarding the force of the person, and the state of the target determined by the target state determination means. Corrective action type determining means for determining the type of corrective action for correcting
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the domestic work of the robot arm determined in advance. A robot arm control device comprising: motion correction means for controlling the robot arm and correcting the motion according to the type.
Information on the operation includes information on the position of the robot arm according to the housework performed by the robot arm, information on force applied to the work surface from the robot arm, information on the direction of the robot arm, The information on at least one of information on the strength of the suction force of the robot arm, speed information on the robot arm, and work non-operation area information which is information on an area where the robot arm is not operated. The control device for the robot arm described in 1.
The information on the operation includes at least force information applied from the robot arm to the work surface according to the housework performed by the robot arm, and information on the strength of the suction force of the robot arm,
The motion correction unit is configured to perform a force control mode in which the robot arm can move in a xyz-axis direction in which a predetermined force is applied to the work surface from the robot arm based on the information related to the motion. The operation before the correction operation according to the force of the person detected by the force detection means and acquired by the information acquisition unit while performing the operation by the robot arm by setting for each axis The robot arm control device according to claim 1, wherein the magnitude or direction of the set force among the information on the robot is corrected.
The information related to the operation includes information on the position of the robot arm, information on the direction of the robot arm, speed information on the robot arm, and a region where no work is performed, according to the housework performed by the robot arm. Non-workable area information that is information about
The motion correction means is configured to apply a force applied to the robot arm from the person to the robot arm while operating in a position control mode for controlling the position of the robot arm based on information on the motion. In response, the impedance detection mode in which the robot arm operates is set for each axis in the xyz axis direction in which the robot arm can move, and the operation is performed while the operation is being performed. The robot arm control device according to claim 1, wherein the operation of the information related to the operation in the impedance control is corrected according to the force of the person acquired by the information acquisition unit.
Control parameter management means for setting a mechanical impedance setting value of the robot arm based on the type of the correction action determined by the correction action type determination means;
The robot arm according to any one of claims 1 to 4, further comprising impedance control means for controlling the mechanical impedance value of the robot arm to the mechanical impedance setting value set by the control parameter management means. Control device.
The impedance control means individually sets mechanical impedance setting values for the six axes of the translation direction and the rotation direction of the hand of the robot arm based on the type of the correction operation,
Further, when correcting the direction of the robot arm of the hand as the type of the correction operation determined by the correction operation type determination unit, the control parameter management means, the rigidity of the robot arm in the direction to be corrected, The robot arm control device according to claim 5, wherein the mechanical impedance setting value is set to be higher than a rigidity of the robot arm in a direction different from the direction to be corrected.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
When the force component in the direction perpendicular to the work surface is equal to or less than a first threshold and the force component in the direction parallel to the work surface is equal to or greater than a second threshold, When the amount of movement of the position of the hand of the robot arm detected by the correction operation type determining means in the direction parallel to the work surface is equal to or greater than a third threshold value, the position of the work surface is determined as the type of correction operation. Decide that it is the type of movement,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to claim 1, wherein the position of the hand of the robot arm is corrected in a direction parallel to the work surface.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is greater than a fourth threshold, the type of the correction operation is determined as the type of movement of the position in the direction perpendicular to the work surface,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination means, The robot arm control device according to claim 1, wherein the position of the hand of the robot arm is corrected in a direction perpendicular to the work surface.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is less than or equal to a fourth threshold and the housework operation is a wiping operation, the type of correction operation is determined to be a correction type of force application ,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to claim 1, wherein a force applied by the robot arm in a direction perpendicular to the work surface is corrected.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction perpendicular to the work surface,
The correction operation type determination means has a force component in a direction perpendicular to the work surface equal to or greater than a first threshold, and the operation at the position of the hand of the robot arm detected by the correction operation type determination means. When the amount of movement in the direction perpendicular to the surface is a fourth threshold value or less and the housework work is a suction work, it is determined that the type of the correction operation is a type of correction of the suction force,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to claim 1, wherein the suction force in a direction perpendicular to the work surface is corrected.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
The correction operation type determination unit is configured to correct the correction when a force component in a direction perpendicular to the work surface is less than a first threshold value and a force component in a direction parallel to the work surface is greater than or equal to a second threshold value. When the movement amount in the direction parallel to the work surface of the position of the hand of the robot arm detected by the action type determining means is less than a third threshold, the correction action type is a speed correction type. Determined that there was
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination means, The robot arm control device according to claim 1, wherein a speed of the position of the hand of the robot arm is corrected in a direction parallel to the work surface.
The correction operation type determination unit measures a displacement amount of the force applied to the robot arm based on the human force applied to the robot arm detected by the force detection unit and acquired by the information acquisition unit, and a measurement result If the displacement amount of the posture component is larger than the displacement amount of the position component, the type of correction operation is determined as the type of posture correction. And
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction operation determined by the correction operation type determination means, The robot arm control device according to claim 1, wherein the posture of the hand of the robot arm is corrected.
The correction operation type determination means detects the amount of movement of the position of the hand of the robot arm in a direction parallel to the work surface,
The correction operation type determination means is configured so that a force applied to the robot arm by the human hand is parallel to the work surface and parallel to the work surface for a certain time detected by the correction operation type determination means. When the amount of movement in the direction is equal to or greater than a threshold, it is determined that the type of correction operation is the type of setting of the work impossible area,
Further, the motion correction means, according to the information on the human force detected by the force detection means and acquired by the information acquisition unit and the type of the correction action determined by the correction action type determination means, The robot arm control device according to claim 1, wherein the work impossible area is set by moving a position of the hand of the robot arm.
The robot according to any one of claims 1 to 4, 7 to 13, further comprising display means for displaying information on the type of the correction operation based on the type of the correction operation determined by the correction operation type determination unit. Arm control device.
A control method of a robot arm that controls the operation of a robot arm and performs housework work while acting on an object of housework work by the robot arm in the home,
Detecting the force of the person acting on the robot arm with force detection means,
Information on the operation including the position of the robot arm in the housework and information on the human force detected by the force detection means are respectively acquired by the information acquisition unit,
The state of the object is determined by the object state determination means,
The motion from the information regarding the motion including the position of the robot arm in the housework work acquired by the information acquisition unit, the information regarding the force of the person, and the state of the target determined by the target state determination means. The corrective action type determining means determines the corrective action type for correcting the human force detected by the force detecting means and acquired by the information acquisition unit during the housework of the robot arm determined in advance. The robot arm control method of controlling the robot arm and correcting the motion by the motion correction means according to the information related to the information and the type of the correction motion determined by the correction motion type determination means.
The robot arm;
A housework robot comprising the robot arm control device according to any one of claims 1 to 4 and 7 to 13, which controls the robot arm.
A control program for a robot arm that controls the operation of the robot arm and performs housework work while acting on the object of housework work by the robot arm in the home,
Information related to the operation including the position of the robot arm in the housework and information detected by the force detection means and acquired by the information acquisition unit and determined by the object state determination means Determining a type of correction operation for correcting the operation from the state of the object by a correction operation type determination unit;
Information regarding the human force detected by the force detection means and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination means during the housework work of the robot arm determined in advance. A robot arm control program for causing a computer to execute a motion correction step of controlling the robot arm and correcting the motion by motion correction means according to a type.
An integrated electronic circuit for controlling a robot arm that controls the operation of the robot arm and performs housework while acting on the object of housework by the robot arm in the home,
Information related to the operation including the position of the robot arm in the housework and information detected by the force detection means and acquired by the information acquisition unit and determined by the object state determination means A correction operation type determining means for determining a type of correction operation for correcting the operation from the state of the object;
Information regarding the human force detected by the force detection unit and acquired by the information acquisition unit and the correction operation determined by the correction operation type determination unit during the housework work of the robot arm determined in advance and the correction operation type determination unit An integrated electronic circuit for controlling a robot arm, comprising: motion correcting means for controlling the robot arm and correcting the motion according to a type.