JP2010035981A - Self-propelled cleaning robot and device, method and program for controlling it - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control device capable of achieving cleaning movement corresponding to a kind of a face to be cleaned to accommodate the movement to surface-direction characteristics of the face when a self-propelled cleaning robot cleans the face. <P>SOLUTION: The kind of a face to be cleaned and surface-direction characteristics of the face 6 are assumed based on a force generated between the face 6 to be cleaned and a cleaning member 8 and from information on a position and a posture of an actuating-handle fixing member 30 of a robot body 10, thus deciding a cleaning track of the body 10. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a self-propelled cleaning robot that performs self-propelled cleaning automatically, a control device and method for the self-propelled cleaning robot, and a control program for the self-propelled cleaning robot.

  In recent years, home-use self-propelled cleaning robots that automatically clean the home have been actively developed. In the conventional cleaning performed by a person, the person often determines the direction of cleaning in consideration of the direction of tatami or flooring. As an example of a robot apparatus that takes such a cleaning method into consideration, when a tatami is selected with a floor selection switch, the amount of light reflected from the floor is detected, and the direction of the texture of the tatami is determined from the time-series change in the amount of light. A self-propelled vacuum cleaner that detects and travels based on the direction of the detected texture has been proposed (see Patent Document 1). In a general vacuum cleaner, the suction brush position is determined according to the floor surface. As an example of a robot device that takes into account such a cleaning method, a light receiving sensor that receives light from the floor surface is provided. A self-propelled vacuum cleaner that realizes more accurate discrimination of the floor material based on the detection result of the sensor and enables fine cleaning is proposed (see Patent Document 2).

JP 2007-319485 A Japanese Patent Laying-Open No. 2005-211364

  However, the conventional technique such as Patent Document 1 has a problem that it does not consider the case of cleaning other than tatami mats. Also, in order to detect the amount of light, if it is assumed that it will be used in the absence of a person, such as a self-propelled cleaning robot, a light source must be provided in the self-propelled cleaning robot, etc. There was a need. In the conventional technology such as Patent Document 2, although the material of the cleaning surface is determined, there is a problem that the cleaning direction is not considered.

  The object of the present invention is made in view of such problems, and when a self-propelled cleaning robot performs cleaning, it detects the force generated between the cleaning surface and the cleaning means, thereby cleaning the object. A self-propelled cleaning robot that can perform cleaning in an appropriate cleaning direction that has been performed by humans by changing the traveling direction of the cleaning means such as a cleaning nozzle or wiping cleaning means according to the material of the surface to be cleaned The present invention provides a control device and method for a self-propelled cleaning robot, and a control program for a self-propelled cleaning robot.

In order to achieve the above object, the present invention is configured as follows.
According to the first aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body arranged on the moving body and moving the moving body by the moving means, In a self-propelled cleaning robot having a cleaning means for cleaning the cleaning surface,
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot that moves the moving body along the cleaning track generated by the track generation unit by the moving unit and cleans the cleaning surface by the cleaning unit is provided.
With such a configuration, it is possible to detect the surface direction characteristic of the surface to be cleaned and control the robot that performs cleaning in the direction along the surface direction characteristic.

According to the eighth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body disposed on the moving body and moving the moving body by the moving means, A control device for a self-propelled cleaning robot comprising a cleaning means for cleaning the cleaning surface,
A force detecting unit detects a force acting on the cleaning unit and the cleaning surface, and based on the force output from the force detection unit, the cleaning unit is brought into contact with the cleaning surface of the floor surface to the cleaning surface. Cleaning direction determining means for determining the direction of cleaning along,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. A control device for a self-propelled cleaning robot is provided.

According to the ninth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body arranged on the moving body and moving the moving body by the moving means, A control method for a self-propelled cleaning robot comprising a cleaning means for cleaning a cleaning surface,
A force detector that detects and outputs a force acting on the cleaning means and the surface to be cleaned;
Based on the force detected by the force detection unit, the cleaning means is brought into contact with the cleaning surface of the floor surface, the cleaning direction determining means for determining the cleaning direction of the surface to be cleaned along the cleaning surface,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means of the cleaning means;
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. A control method for a self-propelled cleaning robot is provided.

According to the tenth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body disposed on the moving body and moving the moving body by the moving means, A control program for a self-propelled cleaning robot comprising a cleaning means for cleaning the cleaning surface,
Computer
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
While functioning as a trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot for causing the moving means to move along the cleaning orbit generated by the orbit generation means and to perform cleaning on the cleaning surface by the cleaning means. Provide a control program.
With such a configuration, it is possible to detect the surface direction characteristic of the surface to be cleaned and control the robot that performs cleaning in the direction along the surface direction characteristic.

  As described above, according to the self-propelled cleaning robot of the present invention, the control device and method for the self-propelled cleaning robot, and the control program for the self-propelled cleaning robot, the self-propelled cleaning is performed while self-propelled. In an automatic cleaning robot, the type of cleaning surface is detected by estimating the type of cleaning surface and the surface direction characteristics of the surface to be cleaned from the force acting on the cleaning means for cleaning the cleaning surface and the cleaning surface for cleaning. Robot control that performs cleaning in a direction along the corresponding surface direction characteristics can be realized.

  DESCRIPTION OF EMBODIMENTS Hereinafter, various embodiments of the present invention will be described before detailed description of embodiments of the present invention with reference to the drawings.

According to the first aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body arranged on the moving body and moving the moving body by the moving means, In a self-propelled cleaning robot having a cleaning means for cleaning the cleaning surface,
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot that moves the moving body along the cleaning track generated by the track generation unit by the moving unit and cleans the cleaning surface by the cleaning unit is provided.
With such a configuration, it is possible to detect the surface direction characteristic of the surface to be cleaned and control the robot that performs cleaning in the direction along the surface direction characteristic.

According to the second aspect of the present invention, the cleaning direction determining means is based on a component that is horizontal and perpendicular to the cleaning surface out of the force detected by the force detection unit. A self-propelled cleaning robot according to a first aspect for determining a direction is provided.
With this configuration, it is possible to realize control of the robot that performs cleaning in the direction along the surface direction characteristics according to the type of the surface to be cleaned.

According to the third aspect of the present invention, a cleaning surface type database having cleaning surface state information of the cleaning surface;
The self-propelled cleaning robot according to the first or second aspect, further comprising a cleaning surface type selection unit that selects a type of the cleaning surface with reference to the cleaning surface type database based on the force detected by the force detection unit. I will provide a.

  According to a fourth aspect of the present invention, the cleaning surface type database includes flooring cleaning surface state information, tatami cleaning surface state information, and carpet cleaning surface state information as the cleaning surface state information. A self-propelled cleaning robot according to the third aspect is provided.

  According to the fifth aspect of the present invention, the cleaning surface type database includes flooring cleaning surface state information, tatami cleaning surface state information, and carpet cleaning surface state information as the cleaning surface state information. The self-propelled cleaning robot according to the third aspect, wherein the cleaning surface type selection means selects the type of the cleaning surface from at least two types of the three types of cleaning surface state information in the cleaning surface type database. provide.

  According to the sixth aspect of the present invention, the cleaning surface type selection means may determine the cleaning surface condition information of the flooring and the cleaning surface of the tatami mat based on the size of the friction coefficient of the cleaning surface. The self-propelled cleaning robot according to the fourth or fifth aspect is provided for selecting any one of the state information and the cleaning surface state information of the carpet.

  According to a seventh aspect of the present invention, in the first to sixth aspects, the trajectory generating means generates a trajectory that changes a region to be cleaned based on the force detected by the force detection unit. Provide a self-propelled cleaning robot.

According to the eighth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body disposed on the moving body and moving the moving body by the moving means, A control device for a self-propelled cleaning robot comprising a cleaning means for cleaning the cleaning surface,
A force detecting unit detects a force acting on the cleaning unit and the cleaning surface, and based on the force output from the force detection unit, the cleaning unit is brought into contact with the cleaning surface of the floor surface to the cleaning surface. Cleaning direction determining means for determining the direction of cleaning along,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. A control device for a self-propelled cleaning robot is provided.

According to the ninth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body arranged on the moving body and moving the moving body by the moving means, A control method for a self-propelled cleaning robot comprising a cleaning means for cleaning a cleaning surface,
A force detector that detects and outputs a force acting on the cleaning means and the surface to be cleaned;
Based on the force detected by the force detection unit, the cleaning means is brought into contact with the cleaning surface of the floor surface, the cleaning direction determining means for determining the cleaning direction of the surface to be cleaned along the cleaning surface,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means of the cleaning means;
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. A control method for a self-propelled cleaning robot is provided.

According to the tenth aspect of the present invention, the moving body, the moving means for moving the moving body on the floor surface, the moving body disposed on the moving body and moving the moving body by the moving means, A control program for a self-propelled cleaning robot comprising a cleaning means for cleaning the cleaning surface,
Computer
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
While functioning as a trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot for causing the moving means to move along the cleaning orbit generated by the orbit generation means and to perform cleaning on the cleaning surface by the cleaning means. Provide a control program.
With such a configuration, it is possible to detect the surface direction characteristic of the surface to be cleaned and control the robot that performs cleaning in the direction along the surface direction characteristic.

(First embodiment)
Hereinafter, a self-propelled cleaning robot, a control device and a control method for the self-propelled cleaning robot, and a control program for the self-propelled cleaning robot according to the first embodiment of the present invention will be described in detail with reference to the drawings. .

First, the configuration of the self-propelled cleaning robot according to the first embodiment of the present invention will be described.
FIG. 1 is a diagram showing an outline of a self-propelled cleaning robot 1 according to the first embodiment of the present invention. In FIG. 1, a self-propelled cleaning robot 1 includes a body part 2 and a robot arm 5 attached to the body part 2, and is placed on a floor surface 9. The mobile body 70 is placed on the floor surface. 9 and a moving means 71 that moves on the front surface of the moving body 70 and moves the moving body 70 by the moving means 71 to a cleaning surface (region or surface to be cleaned) of the floor surface 9. A wiping cleaning member 8 as an example of a cleaning means for cleaning is provided.

  The moving means 71 is, for example, all disposed in the body portion 2, a pair of motors 40 as an example of a driving means for moving the moving body 70, and a pair of motors disposed at the rear portion of the body portion 2. A pair of wheels 3 (right wheel 3 a and left wheel 3 b) that are driven to rotate forward and backward at 40, and an auxiliary wheel 4 that is disposed at the front of the body portion 2.

  The base end of the robot arm 5 is attached to the front end portion of the body portion 2, and the wiping cleaning member 8 is attached to the tip of the robot arm 5 via a force sensor 7 as an example of a force detection unit. .

  The self-propelled cleaning robot 1 is a self-propelled cleaning robot that wipes dirt such as a floor surface, a wall, or a car and wipes and cleans it using a cleaning member 8. In the following description, the case where the floor surface 9 is cleaned will be described as an example.

  2 and 3 are diagrams showing a configuration of the self-propelled cleaning robot 1, and a control device 11 that controls the operation of the self-propelled cleaning robot 1 and a cleaning robot main body that is a control target of the control device 11. FIG. The robot body 10 includes a moving body 70 and moving means 71.

  The control device 11 is configured by a general personal computer as an example, and includes at least a cleaning direction determination unit (an example of a cleaning direction determination unit) 91, a trajectory generation unit (an example of a trajectory generation unit), and an operation. An operation control unit 13 is provided as an example of the control means.

  More specifically, the control device 11 further includes a trajectory generation unit 94 having an operation trajectory database 14 and an operation trajectory generation unit 95, a cleaning surface type database 15, and a cleaning surface type selection unit (an example of a cleaning surface type selection unit). ) 18, an area determination unit 90, and an input / output IF (interface) 16. 2, the operation control unit 13, the trajectory generation unit 94, the cleaning surface type switching unit 19, the region determination unit 90, and the cleaning direction determination unit 91 are configured as a control program 6. Is possible.

  The input / output IF 16 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 control device 11 is connected to a motor driver 17 that drives each link manipulator of the robot arm 5 via an input / output IF 16 as an example of an input unit, and sends a control signal to the motor driver 17.

  By executing the control device 11 that controls the operation of the robot body 10, each joint angle information output from each joint of the robot arm 5 and encoders 21 and 41 (described later) of the pair of body wheels 3, and each Each of the rotation angle information is detected, and further, force information generated at the hand (hand attachment member 30) of the robot arm 5 output from the force detection unit 7 described later is transmitted to the control device 11 through the counter board of the input / output IF 16. The control device 11 calculates a control command value for the rotation operation of each joint portion and the pair of main body wheels 3 based on each joint angle information and each rotation angle information and force information. The calculated control command values are given to the motor driver 17 for driving and controlling each joint portion of the robot arm 5 and the pair of main body wheels 3 through the D / A board of the input / output IF 16, and sent from the motor driver 17. In accordance with each control command value, the motor 20 of each joint part of the robot arm 5 and the two motors 40 of the pair of main body wheels 3 are independently driven. The motor driver 17, the motor 20 of each joint, and the two motors 40 of the pair of main body wheels 3 function as an example of a drive unit. The encoders 21 and 41 function as an example of an angle detection unit that outputs angle information.

  A robot arm 5 is attached to the front surface portion of the body portion 2 via a base portion to constitute a moving body 70. Two wheels 3 are attached to the rear part of the body part 2, and in accordance with a control command value from the motor driver 17, the pair of wheels 3 are rotated in the same forward and reverse directions by the two motors 40 in synchronization with each other. Is possible. Further, the right and left wheels 3a and 3b rotate differently by the two motors 40 according to the control command value from the motor driver 17, so that the body portion 2 can make a turning motion. The pair of wheels 3a and 3b are rotationally controlled so that the cleaning member 8 can move along the target track.

  As an example, the robot arm 5 is a multi-link manipulator having six degrees of freedom, and includes a hand attachment member 30, a forearm link 32 having a wrist portion 31 to which the hand attachment member 30 is attached at a tip 32 a, An upper arm link 33 having a distal end 33a rotatably connected to the base end 32b, and a base 34 fixedly attached to the front surface portion of the body portion 2 are supported by the base end 33b of the upper arm link 33 being rotatably connected. .

As shown in FIG. 3, the wrist portion 31 has three rotational axes of a fourth joint portion 27, a fifth joint portion 28, and a sixth joint portion 29, and a hand attachment member for the forearm link 32. The 30 relative postures (orientations) can be changed. That is, in FIG. 3, the fourth joint portion 27 can change the relative posture around the Ψ _x- axis (vertical axis) of the hand attachment member 30 with respect to the wrist portion 31. The fifth joint portion 28 changes the relative posture of the hand attachment member 30 with respect to the wrist portion 31 around the ψ _y axis (horizontal axis) perpendicular to the ψ _x axis of the fourth joint portion 27. The sixth joint portion 29 changes the relative posture of the hand attachment member 30 with respect to the wrist portion 31 about the axis ψ _z (horizontal axis) in the tip end direction perpendicular to the ψ _x axis of the fourth joint portion 27. Can do. That is, the hand attachment member 30 can rotate independently with respect to the tip 32 a of the forearm link 32 in the three axial directions of the Ψ _x axis, the Ψ _y axis, and the Ψ _z axis. Further, the hand coordinate system 36 in FIG. 2 and these Ψ _x axis, Ψ _y axis, and Ψ _z axis do not always coincide. The hand coordinate system 36 is a virtual one in which the coordinate system obtained by changing the attitude of the absolute coordinate system 35 as described later is translated to the hand side and the origin is moved to the wrist position, whereas the above-described Ψ _x The axis, the Ψ _y- axis, and the Ψ _z- axis are physical rotation axes of the hand mounting member 30 and the three axes are not necessarily orthogonal.

The proximal end 32 b of the forearm link 32 is rotatable around the third joint portion 26 with respect to the distal end 33 a of the upper arm link 33. In the first embodiment, being rotatable about the third joint portion 26 means being rotatable about a horizontal axis parallel to the Ψ _y axis of the fifth joint portion 28. The base end 33 b of the upper arm link 33 is rotatable around the second joint portion 25 with respect to the base portion 34. In the first embodiment, being rotatable around the second joint portion 25 means being rotatable around a horizontal axis parallel to the Ψ _y- axis of the fifth joint portion 28 (that is, the horizontal axis of the third joint portion 26). It means that there is. The upper movable part 34 a of the base part 34 is rotatable around the first joint part 24 with respect to the lower fixed part 34 b of the base part 34. In the first embodiment, being rotatable about the first joint portion 24 means being rotatable about the Z axis of the absolute coordinate system 35.

  As a result, the robot arm 5 constitutes the multi-link manipulator with 6 degrees of freedom so that it can rotate independently around a total of six axes from the first joint part 24 to the sixth joint part 29. .

  Each joint part constituting the rotating part of each shaft has one member of a pair of members (for example, a rotation side member and a support side member supporting the rotation side member) constituting each joint part. And a motor 20 as an example of a rotational drive device that is driven and controlled by a motor driver 17 described later (actually disposed inside each joint portion of the robot arm 5); It includes an encoder 21 (actually disposed inside each joint portion of the robot arm 5) that detects a rotation phase angle (that is, a joint angle) of the rotation shaft. Therefore, the rotation shaft of the motor 12 provided in one member constituting each joint portion is connected to the other member of each joint portion to rotate the rotation shaft forward and backward, thereby causing the other member to move to the other member. It is possible to rotate around each axis with respect to the member.

Reference numeral 35 denotes an absolute coordinate system (XYZ axes) in which the relative positional relationship is fixed with respect to the lower fixed portion 34b of the pedestal 34. Reference numeral 36 denotes a relative positional relationship that is fixed with respect to the absolute coordinate system 35. This is the hand coordinate system. The origin position O _e (x, y, z) of the hand coordinate system 36 viewed from the absolute coordinate system 35 is set as the position of the hand attachment member 30 of the robot arm 5 (that is, the hand position). Then, the posture of the hand coordinate system 36 viewed from the absolute coordinate system 35 is expressed by coordinates (φ, θ, ψ) using the roll angle, the pitch angle, and the yaw angle. The coordinates (φ, θ, ψ) are set as the hand posture of the robot arm 5 (the posture of the hand attachment member 30), and the hand position and posture vector are vectors r = [x, y, z, φ, θ, ψ]. Define ^T. The roll angle, pitch angle, and yaw angle will be described with reference to FIGS. 4A to 4C. First, consider a coordinate system in which the coordinate system is rotated about the Z axis by an angle φ with the Z axis of the absolute coordinate system 35 as the rotation axis. (See FIG. 4A). The coordinate axes at this time are assumed to be [X ′, Y ′, Z]. Next, the coordinate system is rotated around the Y ′ axis by an angle θ using the Y ′ axis as a rotation axis (see FIG. 4B). The coordinate axes at this time are [X ″, Y ′, Z ″]. Finally, this coordinate system is rotated around the X ″ axis by an angle ψ with the X ″ axis as a rotation axis (see FIG. 4C). 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 36, and the hand coordinate system 36, Assume that the orientation vector of the coordinate system 36 is (φ, θ, ψ).

When controlling the hand position, posture, and force of the robot arm 5, the hand position and posture vector r and the hand force vector F generated at the hand described later are used as a target trajectory generation unit described later. target orbit generated by 45, i.e., so that to follow the operation control unit 13 to the tip unit position and orientation target vector r _d and the target force vector F _d.

  The cleaning member 8 is a part for directly wiping a floor surface or a wall surface assuming a mop, a dust cloth, or a sponge, for example. The cleaning member 8 is detachably attached to the hand attachment member 30 via the cleaning member attachment portion 8a, and is placed lightly on the cleaning surface of the floor surface 9 as an example.

The force detection unit 7 detects or estimates a force generated between the hand attachment member 30 of the robot arm 5 and the cleaning member 8 and outputs the information. The forces in the direction of the absolute coordinates (x, y, z) are (F _x , F _y , F _z ), respectively, and the rotational torques about the axes of the absolute coordinates (x, y, z) are (T _x , T _y). , T _z ), and a force generated at the hand is defined as a vector F = [F _x , F _y , F _z , T _x , T _y , T _z ] ^T. As an example, the force detection unit 7 assumes a 6-axis force sensor between the hand attachment member 30 and the cleaning member 8, but when the cleaning direction is limited (for example, FIG. 24), the force detection unit 7 may be 6 axes or less. It doesn't matter. However, means or apparatus capable of detecting at least the force in the direction perpendicular to the cleaning surface (component perpendicular to the cleaning surface) and the force in the traveling direction on the cleaning surface (component horizontal to the cleaning surface). Is required. The output from the force detection unit 7 is output to the operation control unit 13, the cleaning surface type database 15, and the region determination unit 90 via the input / output IF 16.

Further, as the force detection unit 7, the force estimated from the motor current values of the motors 20 respectively disposed at the fourth joint portion 27, the fifth joint portion 28, and the sixth joint portion 29 of the wrist portion 31 without using a sensor. You may make it comprise as an estimation part. In this case, the external force estimated by the force detector 7 is estimated by the following method. The current sensor 17 a of the motor driver 17 measures a current value i = [i ₁ , i ₂ , i ₃ , i ₄ , i ₅ , i ₆ ] ^T flowing through the motor 20 that drives each joint portion of the robot arm 5. . In the force detector 7, the current value i = [i ₁ , i ₂ , i ₃ , i ₄ , i ₅ , i ₆ ] ^T measured by the current sensor of the motor driver 17 is input to the A / D of the input / output IF 9. Captured through the board. Further, the current value q of the joint angle of each joint is taken into the force detector 7 via the counter board of the input / output IF 16, and the joint torque error compensation output τ from the operation controller 13 described later is the force detector. 7 is taken in.

The force detector 7 functions as an observer, and is generated in each joint by external force applied to the robot arm 5 from each information of the current value i, the current value q of the joint angle, and the joint torque error compensation output τ. Torque τ _ext is calculated. Further, the force detection unit 7 converts the equivalent hand external force F _ext after conversion into an equivalent hand external force F _ext at the hand of the robot arm 5 by F _ext = J _v (q) ^−T τ _ext Is output from the force detection unit 7 to the operation control unit 13, the cleaning surface type database 15, and the region determination unit 90. Here, J _v (q) is a Jacobian matrix calculated by the approximate inverse kinematics calculation unit 54 described later.

  FIG. 5 is a diagram showing a detailed configuration of the control device 11. As described above, the control device 11 includes the cleaning direction determination unit 91, the trajectory generation unit 94, and the operation control unit 13, and detects the force acting on the wiping cleaning member 8 and the cleaning surface of the floor surface 9. Based on the force detected by the force detection unit 7 to be output, a cleaning direction determination unit 91 that determines the direction of cleaning along the cleaning surface by bringing the wiping cleaning member 8 into contact with the cleaning surface of the floor surface 9 and a cleaning direction determination unit And a trajectory generator 94 that generates a cleaning trajectory (target trajectory) along the cleaning direction determined by 91, and a movable body along the cleaning trajectory generated by the trajectory generator 94 by a pair of motors 40 and the like. 70 is moved, and control is performed so that the cleaning surface 8 is cleaned with the wiping cleaning member 8.

  More specifically, as described above, the control device 11 further includes the trajectory generation unit 94 including the operation trajectory database 14 and the operation trajectory generation unit 95, the cleaning surface type database 15, and the cleaning surface type selection unit (cleaning surface). An example of the seed selection means) is comprised of a cleaning surface type switching unit 19 having 18, an area determination unit 90, an operation control unit 13, and an input / output IF (interface) 16. Among them, the motion control unit 13 includes a target trajectory generation unit 45, a force error calculation unit 46, a position error calculation unit 47, a position and force control direction selection unit 48, a force error compensation unit 49, and a joint torque conversion. Unit 50, position error compensation unit 51, dynamics compensation unit 52, forward kinematics calculation unit 53, approximate inverse kinematics calculation unit 54, and joint torque command calculation unit 55. The operation control unit 13 outputs a command value for controlling the operation of the robot body 10 from the 10 target trajectories and information on the current position, posture, and force of the robot body 10.

From the robot body 10, the current value (joint angle vector) vector q = [q ₁ , q ₂ , q ₃ , q ₄ , q _{5 of} the joint angle measured by the encoder 21 of each joint of the robot arm _5. , Q ₆ , q _b1 , q _b2 ] ^T are output and taken into the forward kinematics calculation unit 53 of the operation control unit 13 by the counter board of the input / output IF 9. However, q ₁ , q ₂ , q ₃ , q ₄ , q ₅ , q ₆ , q _b1 , and q _b2 are respectively detected by the encoders 21 and 41, the first joint part 24 and the second joint part 25. And the joint angles of the third joint part 26, the fourth joint part 27, the fifth joint part 28, the sixth joint part 29, the right wheel 3a, and the left wheel 3b.

The target trajectory generation unit 45 outputs information from the motion trajectory database 14, force information (current hand force vector) F from the force detection unit 7, and hand position and posture vectors from a forward kinematics calculation unit 53 described later. based on the r, the target trajectory information for realizing the operation of the robot main body 10 as a target, i.e., outputs the hand position and orientation target vector r _d and the target force vector F _d of the robot arm 5. Operation of the robot main body 10 to the target, the position of each point in the object and work each time in advance in accordance with the _{(t = 0, t = t} 1, t = t 2, ···) (r _{d0, r d1, r d2,} ···) and the force _{(F d0, F d1, F} d2, ···) are given to the operation track database 14, the target track generation unit 45, the operation track database 14 Position (r _d0 , r _d1 , r _d2 ,...) And force (F _d0 ) at the respective times (t = 0, t = t ₁ , t = t _{2 ,.} , F _d1 , F _d2 ,...), A hand position and orientation vector r from the forward kinematics calculation unit 53, and a hand force vector F from the force detection unit 7, and polynomial interpolation is used. , Complement the trajectory between each point, the hand position and posture that is the target trajectory Generating a target vector _{r d} and the target force vector _{F d.}

  In the forward kinematics calculation unit 53, the joint angle vector q, which is the current value q of each joint angle measured by the encoder 21 of each joint part of the robot arm 5 of the robot body 10, is displayed on the counter board of the input / output IF 9. Is input via. Then, the forward kinematics calculation unit 53 performs geometrical scientific calculation of conversion from the joint angle vector q of the robot arm 5 of the robot main body unit 10 to the hand position and posture vector r. Therefore, the forward kinematics calculation unit 53 functions as an example of a hand position detection unit that detects the hand position of the robot arm 5 of the robot body 10. The hand position and posture vector r calculated by the forward kinematics calculation unit 53 is output to the target trajectory generation unit 45, the position error calculation unit 47, the cleaning direction determination unit 91, the cleaning surface type selection unit 18, and the region determination unit 90. Is done.

The approximate inverse kinematics calculation unit 54 calculates J _v (q) based on the joint angle vector q measured by the robot arm 5 of the robot body 10. However, J _v (q) is

を満たすヤコビ行列である。ただし、ｖ＝［ｖ_ｘ、ｖ_ｙ、ｖ_ｚ、λ_ｘ、λ_ｙ、λ_ｚ］^Ｔであり、（ｖ_ｘ、ｖ_ｙ、ｖ_ｚ）は手先座標系３６でのロボットアーム５の手先の並進速度、（λ_ｘ、λ_ｙ、λ_ｚ）は手先座標系３６でのロボットアーム５の手先の角速度である。
力誤差計算部４６は、目標軌道生成部４５から出力される目標力ベクトルＦ_ｄと力検出手段７から出力される現在の手先に発生している手先力ベクトルＦとの差である力誤差出力ベクトルＦ_ｅを計算し、位置及び力制御方向選択部４８に力誤差出力ベクトルＦ_ｅを出力する。

Jacobian matrix that satisfies _{However, v} = a _{[v x, v y, v} z, λ x, λ y, λ z] T, (v x, v y, v z) is the hand of the robot arm 5 in the hand coordinate system 36 The translational speed, (λ _x , λ _y , λ _z ) is the angular velocity of the hand of the robot arm 5 in the hand coordinate system 36.
The force error calculation unit 46 outputs a force error that is a difference between the target force vector F _d output from the target trajectory generation unit 45 and the hand force vector F generated at the current hand output from the force detection means 7. The vector F _e is calculated, and the force error output vector F _e is output to the position and force control direction selector 48.

Position error calculation unit 47, the position and the difference between the current hand position and orientation vector r output from the hand position and orientation target vector r _d and the forward kinematics calculation unit 53 outputted from the target track generation unit 45 the attitude error output r _e calculates and outputs the position and orientation error output r _e to the position and force control direction selection unit 48.

  The position and force control direction selection unit 48 separates the direction in which force control is performed and the direction in which position control is performed, and outputs direction correction force error outputs to the force error compensation unit 49 and the position error compensation unit 51, respectively.

（以下、簡略して記載するため、これを「∧Ｆ_ｅ」と記載する。）と方向補正位置及び姿勢誤差出力

(Hereinafter, this is referred to as “∧F _e ” for the sake of brevity.) And the direction correction position and attitude error output

（以下、簡略して記載するため、これを「∧ｒ_ｅ」と記載する。）とを出力する。
例えば、ｘ方向に力制御を行い、かつ、それ以外の方向は位置制御を行う場合には、

(Hereinafter, this is described as “∧r _e ” for the sake of brevity).
For example, when force control is performed in the x direction and position control is performed in other directions,

とすることで、ｘ方向に力制御を行うとともに、それ以外の方向に位置制御を行うことが出来る。

By doing so, it is possible to perform force control in the x direction and position control in other directions.

The force error compensation unit 49 receives the direction correction force error output ∧F _e from the position and force control direction selection unit 48 and controls the robot body 10 so that the direction correction force error output ∧F _e converges to zero. For this reason, a force error compensation output U _{Fe serving} as a control input is calculated, and the force error compensation output U _Fe is output toward the joint torque converter 50.

The joint torque conversion unit 50 converts the force error compensation output U _Fe input from the force error compensation unit 49 into each joint equivalent torque τ _f and directs each joint equivalent torque τ _f to the joint torque command calculation unit 55. Output.

The position error compensation unit 51 receives the direction correction position and posture error output ∧ _e from the position and force control direction selection unit 48, and outputs the position and posture error compensation output U _re to the dynamic compensation unit 52. The

Dynamics compensation unit 52 converts the position and orientation error compensation output U _re inputted from the position error compensation unit 51 in the joint torque tau _r required position error compensation, the joint torque tau _r joint torque command calculated Output to the unit 55.

The joint torque obtained by adding each joint equivalent torque τ _f output from the joint torque conversion unit 50 and each joint torque τ _r required for position error compensation output from the dynamic compensation unit 52 by the joint torque command calculation unit 55. The compensation output τ becomes the output of the motion control unit 13 and is output to the robot body unit 10.

  The joint torque compensation output τ is given as a voltage command value to the motor driver 17 of the robot body 10 via the D / A board of the input / output IF 16, and each joint is driven to rotate forward and backward by the motor 20 of each joint. Thus, the robot body 10 operates.

  With the robot body 10 operating in this manner, the cleaning operation can be performed by moving the cleaning member 8 on the floor surface 9 while performing force control so as to press the cleaning member 8 against the floor surface 9.

  The operation steps of the operation control unit 13 when a track is input from the operation track database 14 to the operation control unit 13 will be described based on the flowchart of FIG. The functions of these operation steps can be executed by the computer as the control program 12 (see FIG. 2).

  Joint angle data (joint variable vector or joint angle vector q) measured by the respective encoders 21 of the joints of the robot arm 5 of the robot body 10 is taken into the control device 11 from the encoder 21 via the input / output IF 16. (Step S21).

Next, based on the joint angle data (joint variable vector or joint angle vector q) taken into the control device 11, the approximate inverse kinematics calculation unit 54 uses the Jacobian matrix J _v necessary for the kinematics calculation of the robot arm 5. Etc. are calculated (step S22).

  Next, in the forward kinematics calculation unit 53, from the joint angle data (joint variable vector or joint angle vector q) from each encoder 21 of the joint part of the robot arm 5 of the robot body part 10, the robot of the robot body part 10 is used. The current hand position and posture vector r of the arm 5 are calculated, and the target trajectory generation unit 45, the position error calculation unit 47, the cleaning direction determination unit 91, the cleaning surface type selection unit 18, and the region determination unit 90 are calculated. The data is output to the motion trajectory generating unit 95 (step S23).

  Next, the force detection unit 7 calculates a hand force vector F generated at the hand of the robot arm 5 based on the output information of the force sensor attached to the robot arm 5 of the robot main body unit 10 to obtain the target trajectory. It outputs to the production | generation part 45, the force error calculation part 46, the cleaning direction determination part 91, the cleaning surface kind selection part 18, the area | region determination part 90, etc. (step S24).

Next, the motion trajectory of the robot body 10 generated by the motion trajectory generation unit 95 described later, the hand position and posture vector r calculated by the forward kinematics calculation unit 53 in step S23, and the force detection unit 7 based on current and hand force vector F that is input from the target trajectory generation section 45 calculates the tip unit position and orientation of the robot arm 5 of the robot main body 10 target vector r _d and the target force vector F _d (step S25 ).

Then, the force error calculation unit 46 calculates a is the difference between the current hand force vector F which is output from the target force vector F _d and the force detecting section 7 from the target track generation unit 45 error output vector F _e, The error output vector _Fe is output to the position and force control direction selection unit 48 (step 26).

Then, the position error calculation unit 47, the hand position and orientation target vector r _d and the current tip unit position and orientation vector r obtained by calculation by the forward kinematical calculation unit 53 in step S23 from the target trajectory generation section 45 of the position and orientation error r _e is the difference calculated, and outputs the position and orientation error r _e to the position and force control direction selection unit 48 (step 27).

Next, the position and force control direction selection unit 48 separates the direction in which force control is performed and the direction in which position control is performed, and outputs the direction correction force error output ∧F _e to the force error compensation unit 49 and the position error compensation unit 51, respectively. and it outputs the direction correction position and the orientation error output ∧r _e (step 28).

Then, the force error compensation unit 49 is a control input for controlling the robot main body 10 so that the position and force direction correction input from the control direction selection unit 48 power error output ∧F _e converges to 0, The force error compensation output U _Fe is calculated, and the force error compensation output U _Fe is output from the force error compensation unit 49 to the joint torque conversion unit 50 (step S29). A specific example of the force error compensator 49 is a PID compensator. The force error compensator 49 is controlled so that the force error converges to 0 by appropriately adjusting the three gains of proportionality, differentiation, and integration, which are constant diagonal matrices.

Next, the joint torque conversion unit 50 converts the force error compensation output U _Fe input from the force error compensation unit 49 into an equivalent joint torque τ _f , and outputs each joint equivalent torque τ _f from the joint torque conversion unit 50. (Step S30).

Then, the position error compensation unit 51 is a control input for controlling the robot arm 5 so that the position and orientation correction position is input from the force control direction selection unit 48 and the orientation error output ∧R _e converges to 0 It calculates the position and orientation error compensation output U _re, and outputs the position and orientation error compensation output U _re from the position error compensating unit 51 on the kinetics compensator 52 (step S31). A specific example of the position error compensator 51 is a PID compensator. The position error compensator 51 is controlled so that the position error converges to 0 by appropriately adjusting three gains of proportionality, differentiation, and integration, which are constant diagonal matrices.

Next, the dynamic compensation unit 52 converts the position and posture error compensation output U _re input from the position error compensation unit 51 into each joint torque τ _r necessary for position error compensation, and outputs each joint torque τ _r . (Step S32).

Next, the equivalent joint torque τ _f calculated in step S30 and the position error compensation joint torque τ _r calculated in step S32 are added by the joint torque command calculation unit 55 to obtain a joint torque compensation output τ. The data is output to the motor driver 17 of the robot arm 10 via the input / output IF 16 (step S33).

  Next, the joint torque compensation output τ is given to the motor driver 17 through the input / output IF 16, and the motor driver 17 changes the amount of current flowing through each motor 20 of the joint portion based on each joint torque compensation output τ. Let Due to the changes in the respective current amounts, rotational movements of the respective joint portions of the robot arm 5 of the robot body portion 10 are generated, and the robot body portion 10 operates (step S34).

  Through the above steps, the robot body 10 can perform an operation based on the trajectory given from the motion trajectory database 14.

  The cleaning surface type switching unit 19 includes a cleaning surface type database 15 and a cleaning surface type selection unit 18, and includes a target trajectory of the robot body 10, a current position, posture, and force of the robot body 10. Select the type of cleaning surface.

  The cleaning surface type database 15 holds information for selecting the type of cleaning surface (an example of cleaning surface state information of the cleaning surface). The type of cleaning surface can be selected based on the hand force vector F input from the force detector 7 and the robot hand position and posture vector r input from the forward kinematics calculator 53. The cleaning surface type database 15 will be described with reference to FIGS.

  First, the schematic diagram of the tatami mat 9a as an example of the kind of cleaning surface is shown in FIG. As shown in FIG. 6A, the tatami mat 9a has a texture 9b. 7A can be detected from the force detection unit 7 when the cleaning member 8 performs a tracing operation in the directions of the five arrows A, A ′, B, B ′, and C shown in FIG. The average friction coefficient of the frictional force generated between the cleaning member 8 and the tatami surface is shown. Here, an arrow A indicates a case where the cleaning member 8 is moved (tracing operation) in a direction along the texture 9b of the tatami 9a, and an arrow A ′ indicates a direction of the weave of the tatami 9a in a direction opposite to the arrow A. The case where the cleaning member 8 is moved in the direction along the eyes 9b (tracing operation) is shown. An arrow B indicates a case where the cleaning member 8 is moved (tracing operation) in a direction perpendicular to the texture 9b of the tatami 9a, and an arrow B ′ indicates a texture 9b of the tatami 9a in a direction opposite to the arrow B. The case where the cleaning member 8 is moved (tracing operation) in a direction crossing at right angles is shown. An arrow C indicates a case where the cleaning member 8 is moved (tracing operation) in a direction that obliquely crosses the texture 9b of the tatami 9a. The friction coefficient can be calculated by the force detection unit 7 by dividing the force in the traveling direction on the cleaning surface among the outputs of the force detection unit 7 during the tracing operation by the normal force of the cleaning surface. When the cleaning member 8 performs a tracing operation on the surface of the tatami 9a in the directions of arrows A and A ′, and when the cleaning member 8 performs a tracing operation in the directions of arrows B and B ′, the cleaning member 8 The friction coefficient of the frictional force generated between the tatami mat and the tatami surface is different, and the direction of the arrow B and the arrow B ′ is larger. Further, the friction coefficient in the other direction (for example, arrow C) is an intermediate value between the friction coefficient in the case of arrow A and the friction coefficient in the case of arrow B. Further, the average friction coefficients when moved in parallel and in opposite directions as indicated by arrows A and A 'and as indicated by arrows B and B' are substantially the same. In FIG. 7B, when the cleaning member 8 performs the tracing operation in each direction of the arrow shown in FIG. 6A, the cleaning member 8 can be detected between the cleaning member 8 and the tatami surface when the force detection unit 7 can detect. The time change of the friction coefficient of the generated friction force is shown. The arrow A, arrow A ', arrow B, arrow B, and arrow C all have little time change during the tracing operation in the same direction. In addition, although the tatami 9a was illustrated here, this information can be applied not only to the tatami 9a but also to a rug or a cushion that is formed by weaving rush (reed grass) which is the same constituent material as the tatami 9a.

  Next, the schematic diagram of the flooring 9c as another example of the kind of cleaning surface is shown in FIG. As shown in FIG. 8, some flooring 9c has a groove 9d in parallel. It can be detected from the force detection unit 7 when the cleaning member 8 performs the tracing operation in the directions of arrows E, E ′, F, F ′, and G on the flooring 9c having the groove 9d as shown in FIG. 9A and 9B show changes in frictional force generated between the cleaning member 8 and the flooring surface with respect to the tracing direction and time. Here, an arrow E indicates a case where the cleaning member 8 is moved (tracing operation) in a direction along the groove 9d of the flooring 9c, and an arrow E ′ indicates a groove 9d of the flooring 9c in a direction opposite to the arrow E. The case where the cleaning member 8 is moved in the direction along (moving) is shown. An arrow F indicates a case in which the cleaning member 8 is moved (tracing operation) in a direction transversely intersecting the groove 9d of the flooring 9c, and an arrow F ′ indicates that the groove 9d of the flooring 9c is perpendicular to the direction opposite to the arrow F. The case where the cleaning member 8 is moved (tracing operation) in the direction crossing is shown. An arrow G indicates a case where the cleaning member 8 is moved (sliding operation) in a direction obliquely crossing the groove 9d of the flooring 9c. As shown in FIG. 9B to FIG. 9D, when the cleaning member 8 performs the tracing operation on the surface of the flooring 9c in the directions of the arrow E and the arrow E ′, the cleaning member 8 does not hit the groove 9d. Always small during operation. In the directions of arrow F, arrow F ', and arrow G, the cleaning member 8 hits the groove 9d, so that the cleaning member 8 is caught and the friction coefficient fluctuates in the middle. In the directions of arrows E and E ′, as shown in FIG. 9B, the friction coefficient does not change in the middle, whereas in the directions of arrows F, F ′, and G, the directions shown in FIGS. 9C and 9D. As shown, the coefficient of friction varies greatly so as to increase midway. For this reason, the average friction coefficient is small in the arrow E and the arrow E ′, and is large in the arrow F, the arrow F ′, and the arrow G. Further, the average friction coefficients when moved in parallel and in opposite directions as indicated by arrows E and E 'and arrows F and F' are substantially the same.

  Next, FIGS. 10A and 10B are schematic views of a carpet 9e having no handle as another example of the type of the cleaning surface. FIG. 10B is a cross-sectional view of the carpet 9e as viewed from the direction A in FIG. 10A. As shown in FIG. 10A, in some regions of the carpet 9e, the hair feet 9f are not vertically but are woven with a direction. The cleaning member 8 and the carpet surface that can be detected by the force detection unit 7 when the cleaning member 8 performs a tracing operation in the directions of arrows H, H ′, I, I ′, and J on the carpet 9e as shown in FIG. 10A. FIG. 11A shows the average friction coefficient of the frictional force generated between the two. Here, the arrow H indicates a case where the cleaning member 8 is moved (tracing operation) in a direction in which the hair foot 9f of the carpet 9e is raised, and an arrow H ′ indicates a hair foot of the carpet 9e in a direction opposite to the arrow H. The case where the cleaning member 8 is moved in the direction in which 9f is laid down (that is, the direction parallel to and opposite to the direction in which the bristle feet 9f are raised) is shown. An arrow I indicates a case in which the cleaning member 8 is moved (tracing operation) in a direction perpendicular to the direction in which the hairs 9f of the carpet 9e are raised or the direction in which the hairs 9f are laid down. In the reverse direction, a case where the cleaning member 8 is moved (tracing operation) in a direction orthogonal to the direction of raising the hair feet 9f of the carpet 9e or the direction of lying down the hair feet 9f is shown. An arrow J indicates a case in which the cleaning member 8 is moved (tracing operation) in a direction that obliquely intersects the direction in which the hair 9f of the carpet 9e is raised or the direction in which the hair 9f is laid down. In the carpet 9e, when the cleaning member 8 performs a tracing operation in a direction in which the hair foot 9f is raised (in the direction of arrow H in FIG. 10A), the friction coefficient is the largest, that is, the direction in which the hair foot 9f is laid down, that is, the hair foot 9f. The friction coefficient becomes the smallest when the cleaning member 8 performs the tracing operation in a direction parallel to and opposite to the direction in which the movement occurs (in the direction of the arrow H ′ in FIG. 10A). In the other direction (in the directions of arrows I, I ', and J in FIG. 10A), an intermediate value of the friction coefficients is taken. In addition, as shown in FIG. 11B, the time change of the friction coefficient during the tracing operation of the cleaning member 8 does not change much when the cleaning member 8 performs the tracing operation in the same direction. Like the coefficient, the direction of arrow H is the largest and the direction of arrow H ′ is the smallest.

Based on the above, an example of the cleaning surface type database 15 is shown in FIG. The cleaning surface type database 15 includes, as an example of cleaning surface state information of the cleaning surface, about the tatami mat, the carpet, and the grooved flooring, the maximum value of the average friction coefficient, and the direction when the average friction coefficient is the maximum value. The difference μ _td between the average friction coefficient in the orthogonal direction, and the difference μ _ta , μ _ja , μ _fa between the maximum value of the average friction coefficient and the average friction coefficient in the opposite direction when the average friction coefficient is the maximum value. , And the differences μ _tp , μ _jp , μ _fp between the maximum and minimum values of the friction coefficient during the tracing operation in the direction in which the average friction coefficient is maximum. However, for carpets and flooring, the difference between the average friction coefficient in the direction in which the average friction coefficient is the maximum value and the direction perpendicular thereto is not specified.

  The cleaning surface type selection unit 18 includes a hand force vector F input from the force detection unit 7, a hand position and posture vector r of the robot body 10 input from the forward kinematics calculation unit 53, and information on the cleaning surface type database 15. Based on the above, the type of cleaning surface is selected. Further, the cleaning surface type selection unit 18 outputs information on the type of the cleaning surface selected by the cleaning surface type selection unit 18 to the cleaning direction determination unit 91 and the operation trajectory generation unit 95.

  The cleaning surface type selection operation (cleaning surface type selection operation) step in the cleaning surface type selection unit 18 will be described with reference to the flowchart of FIG.

First, the hand force vector F input from the force detection unit 7 and the forward kinematics calculation unit 53 input during the search operation (moving operation for obtaining cleaning surface type information) of the robot main body unit 10 described later. based on the hand position and orientation vector r of the robot arm 5 of the robot main body 10, and stored in the cleaning surface type selection unit 18 calculates the direction-of the hand force F _nt tracing operation in the cleaning surface type selection section 18 (See FIG. 14) Further, the force in the direction of the tracing operation is divided by the vertical drag of the tracing surface, and the friction coefficient μ _nt for each tracing operation direction is calculated by the cleaning surface type selecting unit 18 and the cleaning surface type selecting unit 18 Store (see FIG. 15) (step S1).

Specifically, as shown in FIG. 14, the hand forces F _{11 to} F _N1 from direction 1 to direction N at time t = 1 and the hand forces F from direction 1 to direction N at time t = 2. _{12 to} F _N2 , hand forces F _{13 to} F _N3 from direction 1 to direction N at time t = 3,..., Hand forces F _1T from direction 1 to direction N at time t = T and to F _NT is calculated at cleaned surface type selection section 18. Further, as shown in FIG. 15, the friction coefficients μ _{11 to} μ _N1 from the direction 1 to the direction N at the time t = _1, and the friction coefficients μ _{12 to} μ _N2 from the direction 1 to the direction N at the time t = 2. , Friction coefficient μ _{13 to} μ _N3 from direction 1 to direction N at time t = 3, and... Friction coefficient μ _{1T to} μ _NT from direction 1 to direction N at time t = T. This is calculated by the cleaning surface type selection unit 18. Here, the direction refers to each direction of a tracing motion trajectory in a search operation to be described later (see a dotted arrow direction in FIG. 17).

Next, the average value (average friction coefficient) μ _{1 to} μ _N of the friction coefficient for each direction is calculated by the cleaning surface type selection unit 18 and stored (step S2). This, in Figure 15, presented as mean friction coefficients up direction 1 Direction _N μ 1 ~μ _N. At this time, immediately after the start of the tracing operation and immediately before the stop, since the friction coefficient is not stabilized by switching between the static friction and the dynamic friction, the data immediately after the start of the tracing operation and immediately before the stop is not used.

Next, the cleaning surface type selection unit selects the direction in which the average friction coefficient is the maximum value from the average values (average friction coefficient) μ _{1 to} μ _N of the friction coefficient for each direction stored in the cleaning surface type selection unit 18. The cleaning surface type selection unit 18 obtains the difference between the maximum value and the minimum value among the friction coefficients during the tracing operation in the direction in the case of the extracted maximum value (step S3). For example, in FIG. 15, when the cleaning surface type selection unit 18 extracts that the direction when the average friction coefficient is the maximum value is “direction 2”, the direction of “direction 2” which is the direction when the maximum value is extracted. Among the friction coefficients from the time t = 1 to the time t = T during the tracing operation, the cleaning surface type selection unit 18 obtains the difference between the maximum value and the minimum value.

Next, the difference calculated in step S3 is stored in the cleaning surface type database 15, and the maximum value and the minimum value of the friction coefficient during the tracing operation in the direction in which the average friction coefficient in the flooring with grooves is the maximum value. The cleaning surface type selection unit 18 determines whether or not the difference is _{equal to} or greater than μ _fp (step S4).

Hereinafter, in step S4, when the cleaning surface type selection unit 18 determines that the difference calculated in step S3 is equal to or _larger than the difference μ _{fp of} flooring, the process proceeds to step S5. Hereinafter, the operation after step S5 will be described.

  In step S5, based on the information stored in the cleaning surface type selection unit 18, the maximum average friction coefficient, the average friction coefficient during the tracing operation in the direction parallel to and opposite to the direction of the maximum average friction coefficient, and The cleaning surface type selection unit 18 calculates the difference.

Next, the difference calculated in step S5 is stored in the cleaning surface type database 15. The maximum value of the average friction coefficient in the grooved flooring and the average friction in the direction opposite to the direction when the average friction coefficient is the maximum value. The cleaning surface type selection unit 18 determines whether or not the difference from the coefficient _μfa or less (step S6).

Thereafter, in step S6, when the cleaning surface type selection unit 18 determines that the difference calculated in step S5 is equal to or less than the difference _μfa , the process proceeds to step S7.

  In step S 7, the cleaning surface type is determined by the cleaning surface type selection unit 18 as “flooring with grooves”, and information that the surface type is “flooring” is sent from the cleaning surface type selection unit 18 to the cleaning direction. It outputs to the determination part 91 and the operation | movement trajectory generation part 95, and complete | finishes a series of cleaning surface kind selection operation steps. As described above, when the tracing operation is performed in the direction orthogonal to the groove of the flooring, the average friction coefficient becomes the largest, so the direction of the groove of the flooring is the direction orthogonal to the direction of the maximum average friction coefficient.

On the other hand, in step S6, the difference calculated in step S5, if it is determined by the difference mu _fa greater than the cleaned surface type selection section 18 proceeds to step S8.

In step S8, when the difference calculated in step S5 is larger than the difference _μfa , the cleaning surface type cannot be selected even with reference to the information stored in the cleaning surface type database 15, so the cleaning surface type is unknown. If there is, the cleaning surface type selection unit 18 determines that the surface type is unknown, and outputs information indicating that the surface type is unknown to the cleaning direction determination unit 91 and the motion trajectory generation unit 95 from the cleaning surface type selection unit 18. The face type selection operation step is terminated.

Hereinafter, in step S4, when the cleaning surface type selection unit 18 determines that the difference calculated in step S3 is smaller than the difference μ _{fp of} flooring, the process proceeds to step S9. Hereinafter, the operation after step S9 will be described.

  In step S9, based on the information stored in the cleaning surface type selection unit 18, the difference between the maximum average friction coefficient and the average friction coefficient parallel to and opposite to the direction of the maximum average friction coefficient is cleaned. Calculation is performed by the face type selection unit 18.

Next, the difference calculated in step S9 is stored in the cleaning surface type database 15 between the maximum value of the average friction coefficient in the carpet and the average friction coefficient in the direction opposite to the direction when the average friction coefficient is the maximum value. The cleaning surface type selection unit 18 determines whether or not the difference is greater than or equal to _ja (step S10).

The following is a step S10, the difference calculated in step S9 is, if it is determined in cleaned surface type selection section 18 as being the difference mu _niv above, the process proceeds to step S11.

  In step S11, the cleaning surface type selection unit 18 determines that the cleaning surface type is a carpet, and information indicating that the surface type is a carpet is sent from the cleaning surface type selection unit 18 to the cleaning direction determination unit 91 and the motion trajectory generation unit. 95, and a series of cleaning surface type selection operation steps are completed. The direction of the fur (for example, the direction in which the hair feet 9f are raised) is the direction of the maximum average friction coefficient.

On the other hand, in step S10, the difference calculated in step S9 is, if it is determined by the difference mu _niv smaller the cleaned surface type selection section 18 proceeds to step 12.

In step 12, the difference calculated in step 9 is stored in the cleaning surface type database 15, and the maximum value of the average friction coefficient in the tatami mat and the average friction coefficient in the direction opposite to the direction when the average friction coefficient is the maximum value. whether it is less than the difference mu _ta between checked by cleaning surface type selection section 18.

The following is a step 12, the difference calculated in step 9, if it is determined by the difference mu _ta greater than the cleaned surface type selection section 18 proceeds to step S13.

In step S13, when the difference calculated in step 9 is larger than the difference _μta , the cleaning surface type cannot be selected even if referring to the information stored in the cleaning surface type database 15, the cleaning surface type is unknown. If there is, the cleaning surface type selection unit 18 determines that the surface type is unknown, and outputs information indicating that the surface type is unknown from the cleaning surface type selection unit 18 to the cleaning direction determination unit 91 and the motion trajectory generation unit 95. The face type selection operation step is terminated.

On the other hand, if it is determined in step 12 that the difference calculated in step 9 is equal to or less than the difference _μta , the cleaning surface type selection unit 18 proceeds to step S14. Hereinafter, the operation after step S14 will be described.

  In step S14, based on the information stored in the cleaning surface type selection unit 18, the cleaning surface type selection unit 18 determines the difference between the maximum average friction coefficient and the average friction coefficient in the direction orthogonal to the direction of the maximum average friction coefficient. Calculate (step S14). If there is one orthogonal direction, the opposite direction is also orthogonal, so there may be two directions of the maximum average friction coefficient and the direction orthogonal to that direction. The difference is selected by the cleaning surface type selection unit 18.

Next, the difference calculated in step S14 is stored in the cleaning surface type database 15, and the maximum value of the average friction coefficient in the tatami mat and the average friction coefficient in the direction orthogonal to the direction when the average friction coefficient is the maximum value, The cleaning surface type selection unit 18 determines whether or not the difference is greater than or _{equal to μtd} (step 15).

The following is a step 15, the difference calculated in step S14 is, if it is determined in cleaned surface type selection section 18 as being the difference mu _td above, the process proceeds to step S16.

  In step S <b> 16, the cleaning surface type is determined by the tatami mat and the cleaning surface type selecting unit 18, and information indicating that the surface type is tatami is sent to the cleaning direction determining unit 91 and the motion trajectory generating unit 95. To end the series of cleaning surface type selection operation steps. The texture direction of the tatami mat is a direction orthogonal to the direction of the maximum average friction coefficient.

The following is a step 15, the difference calculated in step S14 is, if it is determined by the difference mu _td smaller the cleaned surface type selection section 18 proceeds to step S17.

In step S17, when the difference calculated in step S14 is smaller than the difference _μtd , the cleaning surface type cannot be selected even with reference to the information stored in the cleaning surface type database 15. Therefore, the cleaning surface type is unknown. If there is, the cleaning surface type selection unit 18 determines that the surface type is unknown, and outputs information indicating that the surface type is unknown to the cleaning direction determination unit 91 and the motion trajectory generation unit 95 from the cleaning surface type selection unit 18. The face type selection operation step is terminated.

  As described above, the cleaning surface type selection unit 18 determines the cleaning surface type information based on the information stored in the cleaning surface type database 15, and determines the determined information as the cleaning direction determination unit 91. To the motion trajectory generator 95.

  The cleaning direction determination unit 91 includes a cleaning surface type output from the cleaning surface type switching unit 19, a hand force vector F input from the force detection unit 7, and a robot body unit 10 input from the forward kinematics calculation unit 53. Based on the hand position and posture vector r, a cleaning direction suitable for the type of cleaning surface is determined. The determined cleaning direction is output from the cleaning direction determination unit 91 to the motion trajectory database 14. The “cleaning direction suitable for the type of cleaning surface” determined by the cleaning direction determining unit 91 is a direction that reciprocates along the texture in the tatami mat, and in the carpet, the fur line during the forward operation of the reciprocating operation. This is a direction in which the fur is arranged during the return operation, and in a flooring with a groove, the direction reciprocates in parallel with the groove. In the case of tatami mats, the direction of the texture is determined. In the case of carpets, the direction of fur is determined. This can be determined by the cleaning direction determination unit 91.

  Hereinafter, the cleaning direction determination operation step in the cleaning direction determination unit 91 will be described based on the flowchart of FIG.

First, a hand force vector F input from the force detection unit 7 and a robot input from the forward kinematics calculation unit 53 during a search operation (moving operation for obtaining information on the cleaning direction) of the robot body unit 10 described later. Based on the hand position and posture vector r of the robot arm 5 of the main body 10, the hand force F _nt for each direction of the tracing operation is calculated by the cleaning direction determining unit 91 and stored in the cleaning direction determining unit 91 (FIG. 14). Further, the force in the direction of the tracing operation is divided by the vertical drag of the tracing surface, and the friction coefficient μ _nt for each tracing operation direction is calculated by the cleaning direction determining unit 91 and stored in the cleaning direction determining unit 91 (see FIG. (Refer FIG. 15) (step S61).

  Next, based on the information stored in the cleaning direction determination unit 91, the cleaning direction determination unit 91 calculates the average value of friction coefficients (average friction coefficient) for each direction (step S62). At this time, immediately after the start of the tracing operation and immediately before the stop, since the friction coefficient is not stabilized by switching between the static friction and the dynamic friction, the data immediately after the start of the tracing operation and immediately before the stop is not used.

  Next, based on the information on the cleaning surface type input from the cleaning surface type selection unit 18 to the cleaning direction determination unit 91, the cleaning direction determination unit 91 determines whether or not the cleaning surface type is a carpet (step S63).

  If the cleaning direction determination unit 91 determines in step S63 that the cleaning surface type is carpet, the process proceeds to step S64.

  In step S64, based on the information stored in the cleaning direction determination unit 91, the direction when the average friction coefficient is the maximum value is extracted by the cleaning direction determination unit 91, and the extracted information is extracted from the cleaning direction determination unit 91. Is output to the motion trajectory generation unit 95, and a series of cleaning direction determination operations are terminated. As described above, the direction of the fur is the direction of the maximum average friction coefficient.

  On the other hand, when the cleaning direction determining unit 91 determines in step S63 that the cleaning surface type is other than the carpet, the process proceeds to step S65.

  In step S65, the cleaning direction determination unit 91 determines whether or not the cleaning surface type is flooring based on the cleaning surface type information input from the cleaning surface type selection unit 18 to the cleaning direction determination unit 91.

  If the cleaning direction determining unit 91 determines in step S65 that the cleaning surface type is flooring, the process proceeds to step S66.

  In step S66, based on the information stored in the cleaning direction determination unit 91, the cleaning direction determination unit 91 extracts a direction orthogonal to the direction when the average friction coefficient is the maximum value, and the extracted information is cleaned. The direction determining unit 91 outputs to the motion trajectory generating unit 95, and the series of cleaning direction determining operations is completed. As described above, when the tracing operation is performed in a direction orthogonal to the groove, the average friction coefficient is the largest, so the direction of the groove is a direction orthogonal to the direction of the maximum average friction coefficient.

  On the other hand, when the cleaning direction determination unit 91 determines in step S65 that the cleaning surface type is other than flooring, the process proceeds to step S67.

  In step S67, based on the information on the cleaning surface type input from the cleaning surface type selection unit 18 to the cleaning direction determination unit 91, the direction having the smallest average friction coefficient is extracted by the cleaning direction determination unit 91 and extracted. Information is output from the cleaning direction determination unit 91 to the motion trajectory generation unit 95, and a series of cleaning direction determination operations is completed. The reason why the average friction coefficient is output in the minimum direction is as follows. In the case other than the carpet and the flooring, the case of tatami mat or the case of unknown surface type corresponds, but in the case of tatami mat, the texture direction of the tatami mat is the direction in which the average friction coefficient is the smallest. When the surface type is unknown, it is considered that it is desirable to clean in the direction where the friction, that is, the resistance is small so as not to damage the cleaning surface, so the direction with the minimum average friction coefficient is output.

  As described above, the cleaning direction determination unit 91 is input from the cleaning surface type output from the cleaning surface type switching unit 19, the hand force vector F input from the force detection unit 7, and the forward kinematics calculation unit 53. On the basis of the hand position and posture vector r of the robot main body unit 10, the cleaning direction suitable for the cleaning surface can be determined and the determined information can be output to the motion trajectory generating unit 95.

  The operation trajectory generation unit 95 reads an appropriate cleaning operation trajectory from the operation trajectory database 14 based on the cleaning direction input from the cleaning direction determination unit 91 and the cleaning surface type information input from the cleaning surface type selection unit 18. A trajectory of the hand position and orientation of the robot body 10 is generated and output to the target trajectory generator 45. The cleaning operation generated at the same time is stored. Further, when a search operation to be described later is performed, the search motion trajectory is read from the motion trajectory database 14 and output to the target trajectory generation unit 45.

Here, in the motion trajectory database 14, positions (r _d0 , r _d1 , r _d2 ) for each point at each time (t = 0, t = t ₁ , t = t ₂ ,...) _Are stored in advance. ,...) And force (F _d0 , F _d1 , F _d2 ,...). An example is shown in FIG. FIG. 16 is an example of the operation trajectory database 14 showing the cleaning operation to be performed for each cleaning surface type. In FIG. 16, the plus sign “+” of the cleaning direction means the cleaning direction input from the cleaning direction determination unit 91 to the operation trajectory generation unit 95. A minus sign “−” of the cleaning direction means a direction opposite to the plus direction.

  In the example of FIG. 16, the tatami cleaning is performed by reciprocating along the texture, and there is no difference in the cleaning surface between the back and forth of the reciprocating motion, so the same force (in the example of FIG. 16, 5N) to clean. In the carpet cleaning, first, in the reciprocating motion, cleaning is performed in the direction of raising the hair feet 9f, and then, in the return direction, cleaning is performed in the direction of returning the hair feet. In order to clean the hair while raising the hair, the cleaning member 8 is strongly pressed (10N in the example of FIG. 16) for cleaning, and on the way back, the cleaning member is pressed to adjust the fur. (In the example of FIG. 16, it is reduced to 5N). In flooring cleaning, cleaning is performed by reciprocating in a direction parallel to the groove in the flooring. Since there is no difference in the cleaning surface between the reciprocating movement and the return, the same force (3N in the example of FIG. 16) is applied. Clean. When the type of the cleaning surface is unknown, the cleaning is performed by applying the same force (5N in the example of FIG. 16) on the assumption that there is no difference in the cleaning surface between the reciprocation and the return.

  The motion trajectory database 14 is also provided with a trajectory for a search motion for obtaining information for determining the cleaning surface type and the cleaning surface direction. An example of the search motion trajectory is shown in FIG.

  FIG. 17 is a view of the self-propelled cleaning robot 1 as viewed from above when cleaning the floor surface 9. While pressing the cleaning member 8 against the floor surface 9 as a cleaning surface, the robot body 2 remains stationary and the robot arm 5 is driven to perform a tracing operation with the cleaning member 8 in the direction of the arrow. And the hand force, which is the output from the force detection unit 7 at that time, is input to the cleaning surface type selection unit 18, so that the cleaning surface type selection unit 18 can determine the cleaning surface type and is in operation. The cleaning direction determination unit 91 can determine the type of cleaning surface by inputting the direction and the hand force that is the output from the force detection unit 7 at that time from the force detection unit 7 to the cleaning direction determination unit 91. .

  The region determination unit 90 determines the switching of regions based on the hand force vector F input from the force detection unit 7 and the hand position and posture vector r of the robot body unit 10 input from the forward kinematics calculation unit 53. Then, information on the determination result is output from the region determination unit 90 to the motion trajectory database 14. The region determination unit 90 will be described with reference to FIGS. 18A, 18B, and 19.

  FIG. 18A is an example of how to lay a tatami mat 9a between six tatami mats (how to lay six tatami mats 9a on a rectangular floor of one room). The arrow indicates the locus of the cleaning member 8 of the robot body 10. In the movement trajectories β, β ′, β ″, β ′ ″ along the arrows in the direction parallel to the y-axis in FIG. 18A, the cleaning member 8 is pressed against the cleaning surface to perform wiping, and the direction parallel to the x-axis In the movement trajectories α, α ′, α ″, α ′ ″ along the arrows of the tatami mat 9a, only the cleaning member 8 is moved on the cleaning surface without pressing the cleaning member 8 against the cleaning surface. This shows a state where the cleaning member 8 performs wiping and cleaning along the texture 9b. However, at point A in FIG. 18A, the tatami (texture 9b is parallel to the y-axis) 9a that has been cleaned up to another tatami (texture 9b is parallel to the x-axis) 9a. In another tatami mat 9a, wiping and cleaning are performed along the movement locus of the cleaning member 8 along a direction orthogonal to the texture 9b parallel to the x-axis. FIG. 18B shows the relationship between the change in the friction coefficient during wiping and cleaning, the average value of the friction coefficient during wiping in the positive direction of the y-axis (average friction coefficient μ), and time. When wiping and cleaning is performed on the track of FIG. 18A, the friction coefficient becomes 0 when the cleaning member 8 is moved without pressing, and the friction coefficient increases when the cleaning member 8 is actually cleaning, At point A, the wiper is wiped perpendicularly to the texture 9b, so that the friction coefficient further increases, so the change in the friction coefficient is as shown by the solid line in FIG. 18B. It can be seen that the friction coefficient greatly increases when the solid line in FIG. 18B reaches the time corresponding to the point A near the right end. Further, the average friction coefficient when wiping in the positive direction of the y-axis is indicated by the symbol “x” in FIG. 18B.

  FIG. 19A shows the locus of the carpet 9e and the cleaning member 8 of the robot body 10 during carpet cleaning. In the movement traces β, β ′, β ″ along the arrows in the direction parallel to the y-axis in FIG. 19A, the cleaning member 8 is pressed against the cleaning surface to perform wiping, and along the arrows in the direction parallel to the x-axis. In the movement trajectories α, α ′, α ″, α ′ ″, only the cleaning member 8 is moved on the cleaning surface without pressing the cleaning member 8 against the cleaning surface. It shows a state where wiping and cleaning are performed. However, at point B in FIG. 19A, the cleaning surface type is switched by switching from the carpet 9e to the flooring 9c. FIG. 19B shows the relationship between the change in the friction coefficient during the wiping cleaning, the average value of the friction coefficient during the wiping operation in the positive direction of the y-axis (average friction coefficient μ), and time. When wiping and cleaning is performed on the track of FIG. 19A, the friction coefficient becomes 0 when the cleaning member 8 moves without being pressed, and the friction coefficient increases when the cleaning member 8 is actually cleaning, The coefficient of friction is larger when wiping is performed along the positive direction of the y-axis than when wiping is performed in the negative direction of the y-axis. Furthermore, at the point B, the carpet 9e is finished and the flooring 9c is wiped and cleaned, so the friction coefficient becomes smaller than that during carpet cleaning, so the change in the friction coefficient is as shown by the solid line in FIG. 19B. It can be seen that the friction coefficient is considerably small when the solid line in FIG. 19B reaches the time corresponding to the point B near the right end. The average coefficient of friction when wiping in the positive direction of the y axis is indicated by the symbol “x” in FIG. 19B.

  Therefore, the area determination unit 90 calculates the average friction coefficient during cleaning in the forward direction of the reciprocating movement of the cleaning from the hand force vector F during the actual wiping operation and the hand position and posture vector r of the robot body 10. Are sequentially calculated and stored. Further, when wiping and cleaning in the forward direction, the difference between the current average friction coefficient and the previous average friction coefficient is calculated by the area determination unit 90, and the difference obtained by the calculation exceeds a preset threshold value. If it is determined that the region has changed, the region determination unit 90 determines that the region has changed, and outputs the determined region determination information from the region determination unit 90 to the motion trajectory database 14. When the difference obtained by the calculation is equal to or smaller than a preset threshold value, the region determination unit 90 determines that the region has not changed, and the determined region determination information is transferred from the region determination unit 90 to the motion trajectory database 14. Output.

The above-described cleaning operation of the robot body 10 will be described based on the flowchart of FIG.
First, when the robot body 10 is activated, first, a trajectory of a search operation (for example, a movement operation as indicated by an arrow in FIG. 17) is output from the motion trajectory database 14 to the motion control unit 13 (step S41). .

  Next, based on the trajectory of the search operation, the robot main body unit 10 performs the search operation by driving the robot arm 5 and moving the cleaning member 8 while the robot main body unit 2 remains stationary (step S42). Unlike the cleaning operation, the search operation works in directions that do not follow the direction characteristics of the cleaning surface.Therefore, if the pressing force on the cleaning surface is set smaller than when cleaning, the effect on the cleaning surface can be reduced. It is desirable. The hand position and orientation vector r of the robot arm 5 during the search operation and the hand force vector F are respectively obtained by the forward kinematics calculation unit 53 and the force detection unit 7 of the motion control unit 13.

  Next, the cleaning surface type selection unit 18 performs the cleaning surface type selection operation step of FIG. 13 based on the hand position and posture vector r and the hand force vector F of the robot arm 5 during the search operation of the robot body unit 10. Thus, the cleaning surface type is selected, and the selection result is output to the motion trajectory database 14 and the cleaning direction determination unit 91 (step S43).

  Next, the cleaning direction determination unit 91 is based on the hand position / posture vector r, the hand force vector F, and the cleaning surface type that is the output of the cleaning surface type selection unit 18 during the search operation of the robot body 10 in FIG. By performing the cleaning direction determination operation step, the cleaning direction suitable for the cleaning surface is determined, and the determination result is output to the motion trajectory database 14 (step S44).

  Next, the motion trajectory database 14 outputs the cleaning motion trajectory of the robot body 10 to the target trajectory generation unit 45 of the motion control unit 13 based on the cleaning surface type determination result and the cleaning direction determination result suitable for the cleaning surface. (Step S45).

  Next, the robot body 10 performs the cleaning operation by moving the cleaning member 8 together with the robot body 2 as necessary along the path of the cleaning operation (step S46).

  Next, the region determination unit 90 performs region switching determination based on the hand position / posture vector r and the hand force vector F during the search operation of the robot body 10 (step S47).

  Hereinafter, a case will be described in which the region determination unit 90 determines that the region has not been switched in step S47.

  When the area determination unit 90 determines that the switching has not occurred, the robot main body unit 10 continues cleaning in the cleaning direction according to the direction characteristics of the cleaning surface type by repeating steps S45 to S47. I can do it.

  The following describes the case where the region determination unit 90 determines that a switch has occurred in step S47.

  Next, when the region determining unit 90 determines that the region switching has occurred in step S47, the region determining unit 90 determines that the region switching has occurred from the region determining unit 90 to the motion trajectory generating unit 95. The data is output from the unit 90 (step S48).

  Next, the operation trajectory generation unit 95 generates a trajectory for stopping the cleaning operation. Further, referring to the immediately preceding trajectory stored, the trajectory of the movement to move back to the previous trajectory is generated. The motion trajectory generation unit 95 outputs the above trajectory to the motion control unit 13 (step S49).

  Next, the robot body 10 performs a moving operation based on the trajectory for the movement to the area before switching output to the movement control unit 13, and moves to the area before switching (step S50).

  By repeating the above steps S45 to S50, the robot body 10 can continue cleaning in the cleaning direction corresponding to the direction characteristics of the cleaning surface type.

  FIG. 22 shows an example of the movement trajectory of the cleaning member 8 at the end of the robot arm 5 of the robot body 10 when the robot body 10 moves on the tatami floor based on the flowchart of FIG. The self-propelled cleaning robot 1 is activated at point A in FIG. 22 (the point where the cleaning member 8 is located near the center of the right edge of the right tatami 9a), and the motion trajectory generating unit 95 outputs a search trajectory, The self-propelled cleaning robot 1 performs a search operation. At the same time, the motion trajectory generator 95 stores the output trajectory. (Steps S41 to S42). The cleaning surface type selection unit 18 selects that it is a tatami mat, and the cleaning direction determination unit 91 outputs the texture direction of the tatami mat to the motion trajectory generation unit 95, so that the motion trajectory generation unit 95 ) Outputs the cleaning trajectory along the texture. Thereby, the self-propelled cleaning robot 1 moves while performing cleaning along the texture of the tatami mat (right tatami mat) (steps S43 to S46). In the example of FIG. 22, it gradually proceeds to the left while reciprocating in the direction (vertical direction of FIG. 22) along the texture of the tatami (right tatami). In addition, during this cleaning operation, the region determination unit 90 performs region switching determination. When the self-propelled cleaning robot 1 moves to a place where the tatami is switched while cleaning (point B in FIG. 22 where the tatami is switched from the right tatami to the left tatami), the area determination unit 90 determines that the area is switched (step) S47 to S48). Next, the motion trajectory generation unit 95 generates a trajectory for stopping cleaning. The motion trajectory generation unit 95 further generates a trajectory that returns to the original area based on the immediately previous trajectory that has been stored. The motion trajectory generation unit 95 outputs the created trajectory to the motion control unit 13. Based on the trajectory, the self-propelled cleaning robot 1 moves from point B to point C (steps S49 to S50). Then, again, the motion trajectory database 14 outputs a cleaning trajectory along the texture of the tatami (right tatami), and the self-propelled cleaning robot 1 moves rightward from the C point of the left tatami and moves to the right. Clean the tatami mat.

  The starting and stopping of the robot 1 that performs the above-described cleaning operation will be described with reference to the flowchart of FIG.

First, the robot 1 is activated and performs the cleaning operation described above (step S90).
Next, the power is turned off (step S91). For example, the power may be turned off when a predetermined time is reached by a timer, or the person may generally turn off the power by operating a power button.

  Through the operations in steps S90 to S91, the robot 1 can automatically perform a cleaning operation from the start and stop.

  As described above, according to the cleaning robot 1 according to the first embodiment of the present invention, cleaning is performed by detecting the force generated between the cleaning surface and the cleaning member 8 when cleaning is performed. The self-propelled cleaning robot, the control device and method of the self-propelled cleaning robot, and the self-propelled cleaning robot can be realized by changing the cleaning direction according to the material of the surface. A control program for a traveling cleaning robot can be provided.

  In the first embodiment, the cleaning surface type selected by the cleaning surface switching unit 19 is selected from the three types of carpet, flooring, and tatami. However, it is not always necessary to select the cleaning surface type from the three types. The cleaning surface type may be selected from at least two of them.

(Second Embodiment)
As an example, FIG. 26 shows a flowchart when selecting from two types of flooring and carpet as a second embodiment of the present invention.

  The cleaning surface type selection operation step in the cleaning surface type selection unit 18 when selecting from two types will be described based on the flowchart of FIG. 26 as an example.

First, the hand force vector F input from the force detection unit 7 and the forward kinematics calculation unit 53 input during the above-described search operation of the robot body unit 10 (movement operation for obtaining cleaning surface type information). On the basis of the hand position and posture vector r of the robot arm 5 of the robot main body 10, the hand force F _nt for each direction of the tracing operation is calculated by the cleaning surface type selection unit 18 and stored in the cleaning surface type database 15 ( Further, the force in the direction of the tracing operation is divided by the vertical drag of the tracing surface, and the friction coefficient μ _nt for each tracing operation direction is calculated by the cleaning surface type selection unit 18 and stored in the cleaning surface type database 15. (Refer FIG. 15) (step S71).

Next, the average value (average friction coefficient) μ _{1 to} μ _N of the friction coefficient for each direction is calculated by the cleaning surface type selection unit 18 and stored in the cleaning surface type database 15 (step S72). This, in Figure 15, presented as mean friction coefficients up direction 1 Direction _N μ 1 ~μ _N. At this time, immediately after the start of the tracing operation and immediately before the stop, since the friction coefficient is not stabilized by switching between the static friction and the dynamic friction, the data immediately after the start of the tracing operation and immediately before the stop is not used.

Next, the average friction coefficient is the maximum value among the average values (average friction coefficients) μ _{1 to} μ _N of the friction coefficients for each direction calculated by the cleaning surface type selection unit 18 and stored in the cleaning surface type database 15. The cleaning surface type selection unit 18 extracts the direction of the case, and the cleaning surface type selection unit 18 determines the difference between the maximum value and the minimum value among the friction coefficients during the tracing operation in the direction of the extracted maximum value. Obtained (step S73).

Next, the difference calculated in step S73 is stored in the cleaning surface type database 15, and the maximum value and the minimum value of the friction coefficient during the tracing operation in the direction in which the average friction coefficient in the grooved flooring is the maximum value. The cleaning surface type selection unit 18 determines whether or not the difference is not less than μ _fp (step S74).

Hereinafter, in step S74, if the cleaning surface type selection unit 18 determines that the difference calculated in step S73 is equal to or greater than the difference μ _{fp in} flooring, the process proceeds to step S75. Hereinafter, the operation after step S75 will be described.

  In step S75, based on the information stored in the cleaning surface type database 15, the maximum average friction coefficient and the average friction coefficient during the tracing operation in the direction parallel to and opposite to the direction of the maximum average friction coefficient are calculated. The difference is calculated by the cleaning surface type selection unit 18.

Next, the difference calculated in step S75 is stored in the cleaning surface type database 15, and the maximum value of the average friction coefficient in the flooring with grooves and the average friction in the direction opposite to the direction when the average friction coefficient is the maximum value. The cleaning surface type selection unit 18 determines whether or not the difference from the coefficient is less than or equal to _fa (step S76).

The following is a step S76, the difference calculated in step S75 is, if it is determined in cleaned surface type selection section 18 and the difference is less than mu _fa, the process proceeds to step S77.

  In step S77, the cleaning surface type is determined by the cleaning surface type selection unit 18 to be “flooring with grooves”, and information indicating that the surface type is “flooring” is obtained from the cleaning surface type selection unit 18 as an operation trajectory. It outputs to the database 14 and the cleaning direction determination part 91, and a series of cleaning surface kind selection operation steps are complete | finished. As described above, when the tracing operation is performed in the direction orthogonal to the groove of the flooring, the average friction coefficient becomes the largest, so the direction of the groove of the flooring is the direction orthogonal to the direction of the maximum average friction coefficient.

On the other hand, in step S76, the difference calculated in step S75 is, if it is determined by the difference mu _fa greater than the cleaned surface type selection unit 18, the process proceeds to step S78.

In step S78, when the difference calculated in step S75 is larger than the difference _μfa , the cleaning surface type cannot be selected even with reference to the information stored in the cleaning surface type database 15, and therefore the cleaning surface type is unknown. The information indicating that the surface type is unknown is output from the cleaning surface type selecting unit 18 to the motion trajectory database 14 and the cleaning direction determining unit 91, and the series of cleaning surface type selecting operation steps is completed.

Hereinafter, in step S74, if the cleaning surface type selection unit 18 determines that the difference calculated in step S73 is smaller than the difference μ _{fp of} flooring, the process proceeds to step S79. Hereinafter, operations after step S79 will be described.

  In step S79, based on the information stored in the cleaning surface type database 15, the difference between the maximum average friction coefficient and the average friction coefficient in the direction parallel to and opposite to the direction of the maximum average friction coefficient is determined. The seed selection unit 18 calculates.

Next, the difference calculated in step S79 is stored in the cleaning surface type database 15 between the maximum value of the average friction coefficient in the carpet and the average friction coefficient in the direction opposite to the direction when the average friction coefficient is the maximum value. Whether or not the difference is greater than or equal to _ja is determined by the cleaning surface type selection unit 18 (step S80).

Hereinafter, in step S80, when the cleaning surface type selection unit 18 determines that the difference calculated in step S79 is equal to or _larger than the difference μja, the process proceeds to step S81.

  In step S11, the cleaning surface type selection unit 18 determines that the cleaning surface type is a carpet, and information indicating that the surface type is a carpet is sent from the cleaning surface type selection unit 18 to the motion trajectory database 14 and the cleaning direction determination unit 91. And a series of cleaning surface type selection operation steps are completed (step S81). The direction of fur is the direction of the maximum average friction coefficient.

On the other hand, in step S80, the difference calculated in step S79 is, if it is determined by the difference mu _niv smaller the cleaned surface type selection section 18 proceeds to step 82.

  In step 82, the cleaning surface type selection unit 18 determines that the cleaning surface type is unknown, and information indicating that the surface type is unknown is sent from the cleaning surface type selection unit 18 to the motion trajectory database 14 and the cleaning direction determination unit 91. And a series of cleaning surface type selection operation steps are completed.

  As described above, the cleaning surface type can be selected from at least two types.

  In the first and second embodiments, the cleaning surface switching unit 19 includes the cleaning surface type selection unit 18 and the cleaning surface type database 15 and estimates the cleaning surface type. However, the present invention is not limited to this. As shown in FIG. 27, the robot body 10 has a cleaning surface type selection switch 19S, and a person determines a cleaning surface by the cleaning surface type selection switch 19S and inputs information on the cleaning surface type to the cleaning surface switching unit 19. It may be configured. In this way, the cleaning surface switching unit 19 outputs the cleaning surface type information input from the cleaning surface type selection switch 19S to the operation trajectory database 14 and the cleaning direction determination unit 91, thereby performing a series of cleaning operations. The face type selection operation step can be omitted.

  In the first and second embodiments, after the region determination unit 90 determines that the region is switched (step S48), the cleaning is stopped and moved (step S49). After the determination, the search operation may be performed again to search for a new cleaning direction. If it does in this way, in cleaning of the cleaning surface where an area changes one after another, even if an area changes, an appropriate cleaning direction can be chosen suitably, and cleaning of an appropriate cleaning direction becomes possible.

  In the first and second embodiments, the search operation is performed at the time of start-up, but may be appropriately performed during cleaning. In this way, cleaning in a more appropriate cleaning direction can be realized.

  Further, in the first and second embodiments, as shown in FIG. 1, the suction pump is not limited to the wiping robot provided with the robot arm 5, and for example, as shown in FIG. 23, the suction pump disposed in the robot body 2. By driving the driving circuit 60 and the rotating brush driving device 63 respectively, the suction pump driving circuit 60 drives the suction pump while the rotating brush 64 attached to the suction nozzle 62 rotates by the driving of the rotating brush driving device 63. A suction-type cleaning robot that performs suction by the suction nozzle 66 and collects the dust sucked from the suction nozzle 62 through a suction hose 65 built in the robot arm 5 and collected in a dust bag 61 disposed in the robot body 2. It is good. The rotating brush driving device 63 preferably has a function of pressing the rotating brush 64 toward the cleaning surface of the floor surface 9.

  Moreover, the cleaning robot 1 of the first or second embodiment may be a cleaning robot that does not include the robot arm 5 as shown in FIGS.

  In addition, it can be made to show the effect which each has by combining arbitrary embodiment or arbitrary modifications of the said various embodiment and modifications suitably.

  The present invention relates to a self-propelled cleaning robot, a control device and method for the self-propelled cleaning robot, and a control program for the self-propelled cleaning robot. By changing the direction of travel of the cleaning means such as the cleaning nozzle or the wiping cleaning means according to the material of the cleaning surface to be cleaned by detecting the force generated between It is useful as it can realize cleaning in

The figure which shows the outline | summary of a structure of the self-propelled cleaning robot in 1st Embodiment of this invention. The figure which shows the detailed structure of the robot control apparatus which comprises the said self-propelled cleaning robot in the said 1st Embodiment of this invention, and the cleaning robot main body which is a control object. The figure which shows the detailed structure of the hand of the robot arm of the said self-propelled cleaning robot in the said 1st Embodiment of this invention. Explanatory drawing for demonstrating the roll angle in the hand of the said robot arm controlled by the said robot control apparatus in the said 1st Embodiment of this invention Explanatory drawing for demonstrating the pitch angle in the hand of the said robot arm controlled by the said robot control apparatus in the said 1st Embodiment of this invention Explanatory drawing for demonstrating the yaw angle in the hand of the said robot arm controlled by the said robot control apparatus in the said 1st Embodiment of this invention Control block diagram of the robot control device in the first embodiment of the present invention The figure explaining the texture of the surface of a tatami when the cleaning surface which cleans with the cleaning member of the said robot main-body part controlled by the said robot control apparatus in the said 1st Embodiment of this invention is a tatami. 6 is a graph showing the relationship between the average friction coefficient of the tatami surface and the direction of the tracing operation when the tracing operation in the direction of the arrow in FIG. 6 is performed by the cleaning member of the robot body in the first embodiment of the present invention. The graph which shows the relationship between the friction coefficient of a tatami surface, and time at the time of performing the tracing operation | movement of the arrow direction of FIG. 6 with the cleaning member of the said robot main-body part in the said 1st Embodiment of this invention. The figure explaining the groove | channel of the surface of a flooring, when the cleaning surface cleaned by the cleaning member of the said robot main-body part controlled by the said robot control apparatus in the said 1st Embodiment of this invention is a flooring. The graph which shows the relationship between the average friction coefficient of the flooring surface by the cleaning member of the said robot main-body part controlled by the said robot control apparatus in the said 1st Embodiment of this invention, and the direction of a sliding operation | movement. The graph which shows the relationship between the friction coefficient of a flooring surface, and time when the tracing operation of the direction of the arrow E and arrow E 'of FIG. 8 is performed with the cleaning member of the said robot main-body part in the said 1st Embodiment of this invention. Relationship between the friction coefficient of the flooring surface and the direction of the sliding operation when the cleaning operation of the robot body in the first embodiment of the present invention performs the tracing operation in the directions of arrows F and F ′ in FIG. Graph showing The graph which shows the relationship between the friction coefficient of a flooring surface, and the direction of a sliding operation at the time of performing the tracing operation of the direction of arrow G of FIG. 8 with the cleaning member of the said robot main-body part in the said 1st Embodiment of this invention. The figure explaining the fur of the carpet when the tracing operation | movement of the direction of the arrow is performed with the cleaning member of the said robot main-body part in the said 1st Embodiment of this invention. Sectional view of the carpet seen from the arrow X direction of FIG. 10A The graph explaining the relationship between the average friction coefficient of the carpet surface in the said 1st Embodiment of this invention, and the direction of a sliding operation | movement. The graph explaining the relationship between the friction coefficient of the carpet surface in the said 1st Embodiment of this invention, and time. The figure which shows the cleaning surface kind database in the said 1st Embodiment of this invention. The flowchart showing the operation | movement step of the cleaning surface kind selection part in the said 1st Embodiment of this invention. The figure explaining hand force _Fnt information according to the direction of the tracing operation in the cleaning surface type database used in the self-propelled cleaning robot in the first embodiment of the present invention. The figure explaining the friction coefficient _μnt information according to the direction of the tracing operation in the cleaning surface type database used in the self-propelled cleaning robot in the first embodiment of the present invention. The figure which shows the motion trajectory database used with the said self-propelled cleaning robot in the said 1st Embodiment of this invention. The figure which shows the operation | movement trajectory of search operation in the said 1st Embodiment of this invention. The figure which shows the track | orbit of cleaning by the example of the tatami mat between 6 tatami mats, and the said self-propelled cleaning robot used when explaining the friction coefficient at the time of tatami cleaning in the said 1st Embodiment of this invention FIG. 18A is a graph showing the relationship between the coefficient of friction and time during tatami mat cleaning in the first embodiment of the present invention. Cleaning by the self-propelled cleaning robot in a state where the carpet is ground on a part of the floor of the flooring of one room used for explaining the friction coefficient when cleaning the carpet and flooring in the first embodiment of the present invention. Figure showing the trajectory of In FIG. 19A, the graph which shows the relationship between the coefficient of friction at the time of the carpet and flooring cleaning in the said 1st Embodiment of this invention, and time. The flowchart showing the operation | movement step for performing the robot operation control of the operation control part of the said robot control apparatus in the said 1st Embodiment of this invention. The flowchart showing the operation | movement step of the control part of the control apparatus in embodiment of this invention. The figure which shows the motion trajectory of the robot by the said robot control apparatus in the said 1st Embodiment of this invention. The figure which shows the outline | summary of a structure of the said cleaning robot in the said 1st and 2nd embodiment of this invention. The figure which shows the outline | summary of a structure of the said cleaning robot apparatus in the modification of the said 1st or 2nd embodiment of this invention. The figure which shows the outline | summary of a structure of the said cleaning robot apparatus in the said modification of the said 1st or 2nd embodiment of this invention. The flowchart showing the operation | movement step of the cleaning direction determination part in the said 1st Embodiment of this invention. The flowchart showing the operation | movement step of the cleaning surface kind selection part in the said 2nd Embodiment of this invention. The control block diagram of the robot control apparatus in another modification of the said 1st or 2nd embodiment of this invention. The flowchart showing the operation | movement step from starting of a robot to a stop in another modification of the said 1st or 2nd embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Cleaning robot 2 Body part 3 Wheel 3a Right wheel 3b Left wheel 4 Auxiliary wheel 5 Robot arm 7 Force detection part 8 Cleaning member 8a Cleaning member attachment part 9 Floor surface 10 Robot main body part 11 Controller 12 Control program 13 Operation control part 14 Motion trajectory database 15 Cleaning surface type database 16 Input / output IF
DESCRIPTION OF SYMBOLS 17 Motor driver 18 Cleaning surface type selection part 19 Cleaning surface type switching part 19S Cleaning surface type selection switch 20 Motor 21 Encoder 24 1st joint part 25 2nd joint part 26 3rd joint part 27 4th joint part 28 5th joint part 29 sixth joint part 30 hand attachment member 31 wrist part 32 forearm link 33 upper arm link 34 base part 34a upper movable part 34b lower fixed part 35 absolute coordinate system 36 hand coordinate system 40 motor 41 encoder 45 target trajectory generation part 46 force error Calculation unit 47 Position error calculation unit 48 Position and force control direction selection unit 49 Force error compensation unit 50 Joint torque conversion unit 51 Position error compensation unit 52 Dynamics compensation unit 53 Forward kinematics calculation unit 54 Approximate inverse kinematics calculation unit 55 Joint Torque command calculator 60 Suction pump drive circuit 61 Dust bag 62 Suction nozzle 63 Rotary brush drive device 64 the rotary brush 65 suction hose 66 the suction pump 70 mobile 71 moving means 90 area determination unit 91 cleaning direction decision unit 92 cleaning surface type selection switch 94 trajectory generating unit 95 operating locus generation section

Claims (10)

A moving body, a moving means for moving the moving body on the floor, and a cleaning means arranged on the moving body and cleaning the cleaning surface of the floor by moving the moving body by the moving means In a self-propelled cleaning robot comprising
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot that moves the moving body along the cleaning trajectory generated by the trajectory generating means by the moving means and cleans the cleaning surface by the cleaning means.   The said cleaning direction determination means determines the said cleaning direction of the said cleaning surface based on a component perpendicular | vertical and a component perpendicular | vertical with respect to the said cleaning surface among the forces detected by the said force detection part. Self-propelled cleaning robot. A cleaning surface type database having cleaning surface state information of the cleaning surface;
The self-propelled cleaning robot according to claim 1 or 2, further comprising a cleaning surface type selection unit that selects a type of the cleaning surface with reference to the cleaning surface type database based on the force detected by the force detection unit.   The self-propelled cleaning robot according to claim 3, wherein the cleaning surface type database includes flooring cleaning surface state information, tatami cleaning surface state information, and carpet cleaning surface state information as the cleaning surface state information. .   The cleaning surface type database includes flooring cleaning surface state information, tatami cleaning surface state information, and carpet cleaning surface state information as the cleaning surface state information, and the cleaning surface type selection means includes the cleaning surface state information. The self-propelled cleaning robot according to claim 3, wherein the type of the cleaning surface is selected from at least two types of the three types of cleaning surface state information in the surface type database.   The cleaning surface type selection means, based on the size of the friction coefficient of the cleaning surface, the cleaning surface state information of the flooring, the cleaning surface state information of the tatami mat and the cleaning surface state information of the carpet of the cleaning surface type database The self-propelled cleaning robot according to claim 4 or 5, wherein any one of the above is selected.   The self-propelled cleaning robot according to any one of claims 1 to 6, wherein the trajectory generation unit generates a trajectory for changing a region to be cleaned based on the force detected by the force detection unit. A moving body, a moving means for moving the moving body on the floor, and a cleaning means arranged on the moving body and cleaning the cleaning surface of the floor by moving the moving body by the moving means A control device for a self-propelled cleaning robot comprising:
A force detecting unit detects a force acting on the cleaning unit and the cleaning surface, and based on the force output from the force detection unit, the cleaning unit is brought into contact with the cleaning surface of the floor surface to the cleaning surface. Cleaning direction determining means for determining the direction of cleaning along,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. Control device for self-propelled cleaning robot. A moving body, a moving means for moving the moving body on the floor, and a cleaning means arranged on the moving body and cleaning the cleaning surface of the floor by moving the moving body by the moving means A self-propelled cleaning robot control method comprising:
A force detector that detects and outputs a force acting on the cleaning means and the surface to be cleaned;
Based on the force detected by the force detection unit, the cleaning means is brought into contact with the cleaning surface of the floor surface, the cleaning direction determining means for determining the cleaning direction of the surface to be cleaned along the cleaning surface,
A trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means of the cleaning means;
The moving means controls the self-propelled cleaning robot to move the moving body along the cleaning path generated by the path generation means and to clean the cleaning surface by the cleaning means. Control method of self-propelled cleaning robot. A moving body, a moving means for moving the moving body on the floor, and a cleaning means arranged on the moving body and cleaning the cleaning surface of the floor by moving the moving body by the moving means A control program for a self-propelled cleaning robot comprising:
A force detector that detects and outputs a force acting on the cleaning means and the cleaning surface;
Based on the force detected by the force detector, a cleaning direction determining means for determining a direction in which the cleaning means is brought into contact with the cleaning surface of the floor and cleaning is performed along the cleaning surface;
While functioning as a trajectory generating means for generating a cleaning trajectory along the cleaning direction determined by the cleaning direction determining means,
A self-propelled cleaning robot for causing the moving means to move along the cleaning orbit generated by the orbit generation means and to perform cleaning on the cleaning surface by the cleaning means. Control program.

Images