JPH0423015A - Control system for recognition of object - Google Patents

Abstract

PURPOSE:To improve the measuring accuracy of the position of an object by detecting the collision of a driven matter to the object when a disturbance torque exceeds a prescribed level and then recognizing the object. CONSTITUTION:A servo amplifier 31 drives a servo motor 32 with a torque command. The load of the motor 32 is divided into a load inertia term, a gravity term, a dynamic friction term, etc. The inertia term can be prescribed calculated for each axis. Therefore the torque commands of the terms except the inertia term are approximately constant when the inertia term is excluded from the torque commands. A disturbance torque is suddenly increased when an arm, etc., driven by the motor 32 strikes a work. Based on this increase of the disturbance torque, the work and its position are recognized and a position error is calculated. Then the teaching data is corrected from the position error.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an object recognition control method for recognizing an object by abutting a driven body driven by a servo motor against the object. Regarding the recognition control method.

[Conventional technology]

It would be extremely useful for industrial robots to be able to automatically recognize the type and position of a workpiece and handle it. There are numerous techniques for achieving such goals.

For example, a touch sensor or a force sensor is attached to a driven body driven by a servo motor, that is, the arm of a robot, to recognize the presence or absence of an object or to detect the relative position of the robot and the workpiece. Additionally, a laser sensor is attached to the arm to detect the presence or absence of objects and the relative position of the robot and workpiece. Furthermore, after capturing images of the hand and object using a separately installed video capture device such as a camera, the presence or absence of the object is determined.
Calculate the relative position between the hand and the workpiece. Further, the relative position with respect to the driven body may be calculated using a separately installed three-dimensional position sensor.

Furthermore, the amount of positional deviation of the workpiece is notified to the robot from the force sensor, visual device such as a laser sensor, three-dimensional position sensor, etc. attached to these arms, and the robot corrects the position according to the notified amount of positional deviation. Some devices are configured to automatically correct the teaching program by multiplying the .

In addition, recognition of the object shape may be performed by a worker, or may be performed automatically using a laser sensor, a three-dimensional position sensor, etc.

[Problem to be solved by the invention]

However, the use of touch sensors, force sensors, laser sensors, three-dimensional position sensors, etc. requires installation space and requires data communication with the robot. This makes the robot system complicated and the system cost prohibitive, making it impractical. Therefore, it is far from being a low-cost robot system that can be practically used in actual sites.

Of course, robot lines cannot be automated if the object shape is recognized by dimensional measurements by a worker.

[Means to solve the problem]

In order to solve the above problems, the present invention provides an object recognition control method that recognizes an object by hitting a driven body driven by a servo motor against the object. The disturbance torque expressed as the difference between the simulated torque command value calculated from the load inertia of the motor is calculated every predetermined period, and when the disturbance torque exceeds a separately specified predetermined value, the An object recognition control method is provided, which detects that a driving body hits the object and performs mflc on the object.

[Effect]

A torque command is given to a servo motor that drives a driven body. These torque commands are divided into inertia for accelerating and decelerating the load and other terms. Other than the inertia term, there are gravity terms, recruitment friction terms, etc. These gravitational terms,
The dynamic friction term can be calculated from the position and operating speed of the driven body. As a result, when the driven body hits the object, the dynamic friction term increases and the disturbance torque other than the inertia term increases. When this disturbance torque exceeds a predetermined value, it is detected that the driven body has hit an object, and the object is recognized.

〔Example〕

Hereinafter, one embodiment of the present invention will be described based on the drawings.

FIG. 2 is a diagram showing a horizontally articulated robot for high-speed palletizing. The robot) 1 is mainly composed of a robot body 1a. The robot main body 1a has a servo motor in the Z-axis direction and a mechanical section, and moves the arm 2 and the like up and down. The arm 2 is rotated around the θ axis by a servo motor 3. An arm 4 rotatably coupled to the arm 2 is rotated around the U axis by a servo motor 5. A wrist 6 is provided at the tip of the arm 4, and the wrist 6 rotates around the α axis. Note that various hands for palletizing are used depending on the workpiece, so they are omitted here. The robot 1 hits the workpieces 8a and 8b on the table 7 with its hands, recognizes the workpieces, recognizes the position of the workpieces,
Detects the amount of positional deviation, corrects the operation program, etc. By recognizing or correcting these positions, accurate palletizing can be performed without using special sensors.

The robot 1 also connects the robot control device 10 with a cable l.
Connected by Oa. A teaching operation panel 9 is connected to the robot control device ft210.

FIG. 3 is a partial configuration diagram of the hardware of a robot control device for implementing the present invention. host processor 11
is a processor that controls the entire robot control device 10 according to the system program in the ROM 12. A robot movement command is written from the host processor 11 to the shared RAM 15 at regular intervals. The host processor 11 includes a RAM 13 and a C-ROM for storing the teaching program.
MOSi2 is coupled. The RAM 13 includes a register 13a that stores workpiece position data, which will be described later, and a register 13b that stores taught position correction data.

The DSP (digital signal processor) 21 uses the shared RAM according to the system program in the ROM 22.
15 movement commands are read at regular intervals, the movement command is calculated, the position and speed are detected by the feedback pulse from the pulse coder built into the servo motor 32, and the necessary torque command is sent to the DSL (Digital Servo LSI). 2
4 to the servo amplifier 31. Servo amplifier 31 drives servo motor 32 according to this torque command. The load on the servo motor 32 is divided into load inertia, gravity term, dynamic friction term, etc.

Inertia can be determined and calculated for each axis in advance. Therefore, if inertia is removed from the torque command, the other terms of the torque command are almost constant, and the servo motor 3
2, when the driven arm etc. hits the workpiece, the disturbance torque increases rapidly, and with this increase in disturbance torque, the workpiece is recognized, the workpiece position is recognized, and the amount of positional deviation is measured γ. The teaching data is corrected based on this positional deviation amount. Furthermore, the shape of the workpiece can be recognized from a plurality of position data. Note that necessary calculations and the like are performed using the shared RAM 15.

Note that since the configuration of the control system that controls each servo motor is the same, only the l-axis is shown here.

FIG. 1 is a flowchart of an object position recognition control method according to the present invention. In the figure, the number following S indicates the step number. SI to 56 indicate the processing of the DSP 21.

310 to 513 are processes of the host processor 11, and instructions from the host processor 11 to the DSP 21. 821 to S25 are processes of the main processor,
It is object recognition, position recognition, or processing after position recognition.

[S 1 ] The DSP 21 reads the (SIO) movement command written in the shared RAM 15 by the host processor 11.

[S2] Next, in accordance with this movement command and the feedback signal from the servo motor, a torque command is output, the servo motor 32 is driven, and the arm, which is the driven body, is moved.

[S3] Based on the disturbance torque detection enable signal (Sll) from the host processor 11, it is determined whether or not disturbance torque detection is to be performed, and if it is to be performed, the process advances to S4.

This disturbance torque detection signal is of course sent from the host processor to the DSP 21 via the shared RAM 15.

[S4] Disturbance torque is detected at predetermined time intervals.

The host processor detects the disturbance torque, which is expressed as the difference between the torque command value output to the servo motor driving the robot and the simulated torque command value calculated from the load inertia of the servo motor 32. Detection is performed at predetermined intervals until a possible signal is received. This disturbance torque is compared with the disturbance torque detection level given by the host processor 11 (S12), and when the disturbance torque becomes larger than the detection level, it is determined that the hand has contacted the workpiece.

[S5] Determine whether detection of disturbance torque should be stopped. This is done depending on whether the disturbance torque detection stop signal (S13) is output from the host processor 11.

If the disturbance torque detection signal is output, the process advances to S6; otherwise, the process returns to S4.

[56) The DSP 21 calculates the amount of workpiece pressing from the detected disturbance torque and sends the shared R to the main processor 11.
Notification will be made on AM15.

[521] The main processor 11 reads the coordinate values of the robot control device at this time and stores them in the register 13a.

C822] The amount of positional deviation of the workpiece 8a is calculated from the coordinate values at this time and the teaching program. Further, the amount of positional deviation can be calculated from the difference between the disturbance torque when the workpiece 8a is in the standard position and the actual disturbance torque. This requires determining the relationship between the amount of positional deviation and the disturbance torque in advance.

These positional deviation amounts are stored in the register 13 as position correction amounts.
Write in b.

[323] The host processor 11 uses the position correction amount written in the register 13b to perform position correction for the separately defined posture in the operation program.

[324:] In the operation program after position correction, work 8
Work in an accurate posture relative to a.

Calculate the position correction amount of C82530 bot.

Furthermore, by measuring a plurality of positions of the workpiece 8a, etc., the robot can recognize the three-dimensional shape and measure the dimensions of the workpiece. That is, it is possible to determine whether the work is the work 8a or the work 8b without using a special sensor or the like.

Disturbance torque has the characteristic of varying depending on the operating speed of the driven object, so in order to make the above function more effective, the detection level, detection start, and stop of disturbance torque can be set arbitrarily in the robot's operation program. By doing so, it is possible to improve the accuracy of position measurement, and it is possible to solve problems such as activation of the object presence/absence recognition function at unnecessary positions.

〔Effect of the invention〕

As explained above, in the present invention, the presence or absence of an object, its position, etc. can be recognized by fluctuations in disturbance torque, so a special sensor is not required, and a robot system can be configured with a simple configuration.

[Brief explanation of drawings]

Fig. 1 is a flowchart of the object position recognition control method of the present invention, Fig. 2 is a diagram showing a horizontally articulated robot for high-speed palletizing, and Fig. 3 is a diagram of the hardware of the robot control device for implementing the present invention. FIG. 1 Robot 2 Arm 4 Arm 6 Wrist b Table work work robot control device host processor OM AM-MO3 Shared RAM DSP <Digital processor) OM AM DSL (Digital servo I) Servo amplifier servo motor signal block LS

Claims (7)

[Claims] (1) In an object recognition control method that recognizes an object by bumping a driven object driven by a servo motor against the object, the torque command value output to the servo motor and the load inertia of the servo motor are calculated. Disturbance torque expressed as the difference between the simulated torque command value and What is claimed is: 1. An object recognition control method, comprising: detecting that the object is detected, and recognizing the object. (2) The object recognition control method according to claim 1, wherein the position of the object is measured by reading the coordinate values of the robot when the object is recognized. (3) The object recognition control method according to claim 2, wherein the amount of positional deviation of the object is calculated from the coordinate values of the robot when the object is recognized. (4) Object recognition according to claim 1, wherein the object is recognized and the amount of positional deviation of the object is detected based on a difference in disturbance torque when the driven body is pressed against the target object. control method. (5) The object according to claim 3 or 4, wherein the object calculates its own posture at a separately specified relative position from the amount of positional deviation, and automatically performs positional correction from the taught position. recognition control method. (6) The object recognition control method according to claim 2, wherein a plurality of pieces of positional data of the object are obtained, and shape recognition and dimension measurement of the workpiece are performed from the positional data. (7) The object recognition control method according to claim 1, wherein the detection level of disturbance torque, a detection start command, and a detection stop command are set in the operation program of the robot.