JPH02121007A - Controlling system for teaching robot - Google Patents

Abstract

PURPOSE:To lighten the load of a hand part by using a connection point between a force sensing sensor part and the hand part as a point to be held by a worker for teaching. CONSTITUTION:In the case to teach a robot 10 in work, a connection spot 1a between the force sensing sensor part 4 provided with a force detecting means 12 for teaching and the hand part 3 is used as the point E to be held by the worker, and the worker instructs as holding said connection spot 1a. Force impressed by the worker is detected by the force detecting means 12 as making the point E the point of application of the force, and the detected force and its direction at the point E are outputted. Then, at the time of reproduction, they are reproduced as the action of the hand part 3. Thus, the load of the hand part 3 can be lightened at the time of instruction.

Description

[Detailed description of the invention] [Table of contents] Overview Industrial field of application Prior art (Figure 2) Means for solving the problem to be solved by the invention (Figures 1 and 2) Working examples (Figures 2 to 9) Effects of the invention [Summary] Regarding a robot teaching control method in which a manipulator is manually operated to teach a robot work, the present invention relates to a robot teaching control method in which a manipulator is manually operated to teach a robot work. A force detection means is provided in the force sensor section and detects the force applied to the manipulator, with the purpose of reducing the load on the hand section by using it as a point held by the worker for the purpose of the operation. The force detected by the force detection means is inputted, the displacement caused by the operation of the manipulator is inputted from the position detection means provided in the drive means of the manipulator, and parameters are inputted to determine the reference point coordinate system. force processing means for outputting the force of;
Jacobian matrix calculation means for calculating a transformation matrix for obtaining control amount information from the transformation matrix input from the force processing means and the Jacobian matrix at the position of the force detection means; a position processing means for outputting the current position of the hand section or the connection point, a current position input from the position processing means, a transformation matrix input from the Jacobian matrix calculation means, and a transformation matrix input from the force processing means. control amount information generation means that outputs control amount information on the hand section or the connection point from the force in the reference point coordinate system; I decided to play it back as part of the action.

[Industrial application field]

The present invention relates to a robot teaching control method for teaching robot work by manually operating a manipulator.

[Conventional technology]

As shown by the broken line in FIG. 2, the conventional robot teaching control system, when performing direct teaching, has a hand part 3 at the tip of the joint link part 2 of the manipulator 1, and a hand part 3 attached to the hand part 3. When an operator holds a robot equipped with a force sensor unit 4 that detects the applied force and outputs the movement of the hand unit 3 to the control device side, and manually moves the manipulator 1, the hand unit 3 The force sensor unit 4 detects the applied force, and the robot controller (
(not shown) to input that information. During work, the joint link section 2 is moved in the direction of the force input by the robot controller to reproduce the motion of the hand section 3.

In order to reproduce the movement of the hand section 3 by the robot controller, the force detected by the force sensor section 4 is converted into a force acting on the gripping point H, a movement speed is generated, and a Jacobian matrix regarding the gripping point H is roughly calculated. Processing is performed to generate joint velocity or joint torque of the manipulator 1 by calculation.

[Problem to be solved by the invention]

In the conventional robot teaching control system described above, the operator holds the grip point H of the hand part 3 of the robot manipulator 1 and applies force in the movement direction to move the hand part 3 to the teaching point. This results in a load being applied to the fingers of No. 3 more than necessary, which is unfavorable in terms of strength, and also poses a problem in that it is impossible to hold the object while gripping it.

The present invention has been made in view of the above-mentioned problems, and the technical problem set for the purpose of solving the problem is that a worker has a connection point between a force sensor section and a hand section for teaching. It is an object of the present invention to provide a teaching control system for a mouth pot that can reduce the load on the hand section by using it as a point.

[Means to solve the problem]

As a specific means for solving the above-mentioned problems, the present invention provides a robot teaching control system in which a hand portion 3 is attached to the tip of a manipulator 1, as shown in FIGS. 1 and 2. The robot 10 has a connection point 1a between the hand section 3 and the force sensor section 4, and includes a force detecting means 12 provided in the force sensor section 4 and detecting a force applied to the manipulator 1; Detection means 1
2 inputs the force detected by the position detecting means 1 provided on the driving means 13 of the manipulator 1.
3a, force processing means 14 inputs the displacement caused by the operation of the manipulator 1, further inputs the parameters at each position on the teaching trajectory, and outputs the force in the reference point coordinate system at each position on the teaching trajectory. , a Jacobian matrix calculation means 15 for calculating a transformation matrix for obtaining control amount information from the transformation matrix input from the force processing means 14 and the Jacobian matrix at the position of the force detection means 12, and the position detection means 13a. a position processing means 16 that outputs the current position of the hand section 3 or the connection point 1a based on the displacement inputted from the position processing means 16; The control amount information generation means 17 outputs control amount information at the hand section 3 or the connection point 1a from the transformation matrix and the force in the reference point coordinate system inputted from the force processing means 14, When teaching the work, the connection point 1a is manually operated, and when the work is reproduced, the movement of the hand section 3 is reproduced.

[Effect]

With the above configuration, when teaching the robot 10 a work, the present invention uses the connection point 1a between the force sensor section 4 having the force detection means 12 and the hand section 3 as a point E held by the worker for the teaching. When the operator holds the connection point 1a and teaches, the force applied by the operator is detected by the force detection means 12 with point E as the point of force application, and the force at the detected point E and its direction are output, and the robot 10 When performing a work, the force processing means 14 processes the teaching force F detected by the force detection means 12 with respect to each point on the teaching trajectory drawn by the point of application H of the hand section 3, and the position processing means 16 processes the position of the hand part 3, that is, the joint displacement θ of the manipulator 1 detected on the drive device side, and the Jacobian matrix calculation means 15 inputting the processing results calculates a transformation matrix for coordinate transformation, and the result is If the subscript K is to be displayed as point E or point H by the controlled variable information generating means 17 which has input +PK) as a parameter to obtain drive device control (signal) information such as the joint speed or joint torque of the manipulator 1 necessary for the work operation in the hand section 3, and then control the drive device 13 based on the result. The force in the hand section 3 during work and the corrected movement in the direction are controlled by the manipulator 1.
have it executed.

〔Example 〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS As an embodiment of the present invention, an articulated robot controlled by a microcomputer will be illustrated and explained below.

The device configuration of the articulated robot is as shown in FIG. 3, in which the robot body 20 includes a manipulator 1, a servo motor 22 that operates the manipulator 1, a power amplifier 23 that supplies power to the servo motor 22, A tachometer 24 and an encoder 25 that detect the movement of the servo motor 22, and a force sensor 26 that detects the force applied to the manipulator 1 are provided. The robot controller 30 includes a DA converter 31 that supplies power to the power amplifier 23, and an AD converter 32 and an AD converter that converts the output of the tachometer 24 and encoder 25 into digital signals.
a converter 33, an AD converter 34 that converts the output of the force sensor 26 into a digital signal, and each of these converters 31, 3.
A CPU 36, a RAM 37, a control arithmetic unit 38, and a communication interface 39 which exchange data via the internal bus 35 and the internal bus 35 are provided.

A communication interface 39 provided in the robot controller 30 enables data communication with a host computer 42 and a teaching device 43 via an external bus 41, and the host computer 42 and teaching device 43 communicate data via an external bus 44. Data can be exchanged between the file 45 and the keyboard 46.

In the case of a speed control type multi-articulation robot, the teaching control method for a robot with such a device configuration includes a force sensor 26 on the robot body 20 side, as shown in FIGS. 2 and 4. a force detection means 52 that detects the force applied to the manipulator 1 and outputs a force SF expressed in a reference coordinate system at that position; and a drive means 53 that has a servo motor 22 and operates the manipulator 1. The drive means 53 is provided with a tachometer 24 and an encoder 25, and the current position RX of the connection point 1a between the hand section 3 and the force sensor section 4 as the manipulator 1 moves;
: and a position detection means 53a that outputs:.

The control calculation unit 38 side of the robot controller 30 receives the detected force sFs from the force detection means 52, and the position detection means 53a.
Force processing that inputs the current position RXE of the connection point 1a from the host computer 42 side, the target pressing force RFo and the parameters at that point, and outputs the control force F or the conversion matrix TR/K + UK/S at that point. A means 62 is provided to connect the 9 position detection means 53a to the connection point 1.
A position processing means 63 is provided for inputting the current position RX6 of a and the target position RXO from the host computer 42 side and outputting the control position X at that point, and the force processing means 62
input UK/s into the transformation matrix TR/ to obtain the inverse matrix of the Jacobian matrix Js at the position of the force detection means 52, the inverse matrix of the transposed matrix of the transformation matrix UK/8, and the transformation matrix TR
, and also calculates the product of these to calculate and output an inverse Jacobian matrix J-1.
3 and the output of the Jacobian matrix calculation means 64,
A control amount generating means 65 is provided which calculates the joint velocity at that time and outputs it to the drive device 53 side.

According to the embodiment configured in this way, the manipulator 1
As shown in FIG. 2, the robot has a hand section 3 and a connecting point 1a between the hand section 3 and the force sensor section 4 at the tip of the robot. The force processing means 62, position processing means 63, axis and matrix calculation means 64, and control amount generation means 65 provided in the control calculation section 38 of the controller 30 calculate , the joint displacement θ of the manipulator and the teaching force F detected by the force sensor are input, and the homogeneous transformation element of the three-dimensional coordinate transformation (nK + OK + aK + PK) (here, the subscript indicates point E or point H It becomes possible to process an arithmetic expression whose parameters are (display), and obtain drive device control (signal) information such as the position and speed of each joint necessary for the work movement of the hand section 3. That is, as shown in FIG. 5, the coordinate system (H;
The origin H of HY, Z, ) is taken between the two fingertips for grasping the object, and the vector with the calibration point S of the force sensor unit 4 as the base point is PH= ((PH), (PH)y (PR), )
”

Here T is the transposed matrix.

The three unit vectors representing the posture of the hand section are determined as follows.

The X vector is taken in the direction in which the hand approaches the object, and is called an approach vector and is expressed as n.

The X vector is taken in the direction from one fingertip to the other fingertip, and is called a direction vector and is expressed as 0. the last Z
The vectors are taken so that three vectors form a right-handed system, are called normal vectors, and are denoted by a, and are given by the following equation.

a=nX.

If determined as above, the transformation (4x4 transformation matrix) AH that displays the position and direction of the hand part is the parameter that is the homogeneous transformation element of the three-dimensional coordinate transformation from point 8 to point H. (nH, OH, aH.

PH).

Vector P from 8 points in the sensor coordinate system (S; Xs YS zs) to point E in the teaching coordinate system (E; XE YE ZE). The homogeneous transformation matrix AE for is defined by the following equation, and is completely determined by giving the values of these 16 elements. Therefore, the values for each point on the teaching locus that changes due to the teaching operation are sequentially given.

As shown in FIG. 6, a calibration point S (hereinafter referred to as point S) of the force sensor section 4, a point of application of the teaching force E (hereinafter referred to as point E),
and the point of action H (hereinafter referred to as H point) of the hand part, the parameters at point E are 3 from point 8 to point E.
Homogeneous transformation element of dimensional coordinate transformation (nE + OE + aE
+ PE ), at point H, for example, as shown in Figure 2.
In the case of a 6-joint manipulator, the robot reference point OR
The homogeneous transformation matrix T from to 8 points is given by the following equation, where As is the homogeneous transformation matrix at 8 points.

T = A4 A2 A3 A4 A3As As
”・(4) Here Al (i=1.2...,6)
is a homogeneous transformation matrix that relates the coordinate system of link i to the coordinate system of link i-1. For example, in the case of i = 1 in the manipulator with the joint configuration shown in Figure 2, it is given by (6) .

Using this homogeneous transformation matrix T, the reference point coordinate system (OR;
Hect representing the current position RxE of point E in XRYRZyt ) IiE = (EXEyE2)
3×3) matrix TPE is given by the following equation.

to 1 Koni, It is.

The force EF acting on point E is expressed as a force sFs equivalent to when the points of action are changed to eight points, and is given by the following equation.

EF,=U,...SFs (10) Here, from this matrix Tp0, the force EF6 acting on point E expressed in the teaching coordinate system is expressed as The force RFE is obtained from the following equation.

RF E = TR/E EF E ・・・・・・
...... (8).

By setting SF as the output of the force sensor, (8) to (1
From equation 1), the force EF6 at the point of application E is determined as the force RFo in the reference point coordinate system using the following equation.

RFE =TR/E UE/S SFs ”””(
12) The calculation process to obtain the force RF expressed in the reference point coordinate system is executed by the force processing means 62, and the force RFE obtained as a result of the calculation and the transformation matrix T obtained during the calculation are
R, and UE, are stored.

Assuming that each axis of the hand coordinate system (H; XHYHZH) is a counterclockwise rotation angle with respect to each axis of the reference point coordinate system (OR;
, the rotation angle around the YR axis is β□, and the rotation angle around the ZR axis is γ
. Then, the current position RXH of the hand section 3 is RX, = (hXhyh2α□β8γH)"-(13)
is given by

The current position RXH of the hand section 3 is determined by the position processing means 63.
The manipulator 1 inputted from the position detection means 53a side in
It is calculated from data regarding the displacement θ.

If the articulated robot uses a speed control system, the speeds of the eight points 8■5 and the joint speed θ are expressed in the sensor coordinate system. Which relationship is given by the following equation using the Jacobian matrix Js at the eight points.

5vs=JSθ0 ・・・・・・・・・・・・ (1
4) Here, sy3 is equivalent to the speed EVE at point E, so SVs = (UE/ST) −” EVE ”” (1
5). In addition, ``If VE is expressed as the speed RVE expressed in the reference point coordinate system, EVE = (TR/E) -'' RVE ''
(16).

From equations (14) to (16), joint velocity θ. is given by the following equation.

θo = JS-"(UE/5T)-'(TR/E)
-"RVE" (17) Here, the speed RvE is RvE=Cf RFE (18) for the feedback gain C, which will be described later.
is given by

Therefore, during teaching work, the parameter is set to the value at point E, and the moving speed RVE for the force RF acting on point E is
and joint velocity θ. is generated based on the moving speed at point E, so the manipulator learns the desired movement of the worker who holds point E and gives instructions.

Then, at the end of the teaching, the work reproduction point as a learning result is the grasping point H of the hand part, so regarding the parameters,
Homogeneous transformation element (n
r, + OE + ar: + P E ) as a homogeneous transformation element (nHl
oHI aHI PH).

This replacement is performed outside the robot controller 30,
The result is inputted to the force processing means 62, and the subsequent processing in each means is performed for point H in the same manner as for point E.

The homogeneous transformation matrix An for the vector Po from the 8 points in the sensor coordinate system to the H point in the hand coordinate system is given by the following equation, and for the homogeneous transformation matrix T from the robot reference point OR to the H point, it is defined as The vector h=(hXhyh, )T and the (3X3) matrix TpH are calculated.

Current position RX of the hand determined from this h and TPH
H is RX, = (hXhyh, α□β□γH)T...(
21).

Also, the pressing force RFH at the hand part is RFH=T
R/HUH/S sFs ””...(22),
Joint velocity θ. is θo = J S-”(UH/5T)-”(TR/H)
−1RVH” (23).

Here, the speed RvH is the moving speed "vo" in the direction of the target point input from the host computer 42 side, the speed increment "v" on the pressing side as a force control factor, and the speed increment on the pulling side as a position control factor. It is given as the sum of RV.

The speed increment Rvf is given by the following equation using the target pressing force RFo input from the host computer 42 side.

Here, F is determined by the force processing means 62, the result is inputted to the controlled variable generating means 65, and the velocity increment "Vf" is determined by the controlled variable generating means 65.

The speed increment Rvp uses the target position RXo input from the host computer 42 side, and if ■ is a unit matrix, then
It is given by the following formula.

Here, X is determined by the position processing means 63, the result is input to the controlled variable generating means 65, and the speed increment Rvp is determined by the controlled variable generating means 65.

The coordinate transformation R is is given by

The selection matrix Sf is given by: In this case, in general, pressing, such as tightening a screw or tightening a can lid, applies torque around the axis of the direction, so S=1 is set when torque is applied, and S=0 when no torque is applied.

The feedback gain Cf is expressed as follows with respect to the reference point coordinate system: is given by

In addition, the position feedback gain Cp is similarly (step 71), and is given by

The joint velocity θ given in this way. is calculated by the control amount generating means 65.

In this manner, during work reproduction, control is performed as the movement of the hand section 3 at the gripping point H.

Parameter replacement performed outside the robot controller 30 is performed by the host computer 42 or the teaching device 4.
It will be executed in 3rd place.

If the replacement is performed on the host computer 42, the seventh
As shown in the figure, the host computer 42 is connected to the teaching device 4.
When the teaching start flag is received from 3, S(=diag(1, 1, 1, 1, 1, 1)
...(30)...(31) (Step 72) is executed to transfer the parameters at point E to the address where they are stored. after that,
Check whether the teaching end flag has arrived from the teaching device 43 (step 73). If not, wait for it to come, and if the teaching end flag has been received, the robot controller 30
(32) (step 74) is executed for the RAM 37, and the parameters at point H are newly transferred to the address where they are stored.

In this way, by performing parameter replacement outside the robot controller 30 when the need arises, it becomes possible to efficiently perform the processing in three days in the control calculation section.

In the case of tracing motion, a moving speed RVo is given based on the tracing direction instead of the teaching direction, and the tracing coordinate system x' y
If 'z' has the same definition as the hand coordinate system, it can be the X' direction force control direction, Y' and Z' force direction position control directions, Cf and Cp are the force and position feedback gains, and R Transformation matrix to horizontal coordinate system,
Sf is controlled in the same manner as above as a selection matrix of force control directions in the contour coordinate system. 'Another aspect of the teaching control method in this speed control multi-articulation robot can be configured to be easily changed to an analog control method, as shown in FIG.

In this configuration, the force processing means 81 calculates the force RFH,
Using the intermediate results, the Jacobian matrix calculation means 82 calculates the inverse Jacobian matrix J-1, and the force processing means 81 outputs the force RFH.
is input, and the force control factor calculating means 83 calculates the speed increment Rvf. Then, the position processing means 84 calculates the current position RXH of the hand section 3, and upon inputting the current position RXH, the position control factor calculation means 85 calculates the speed increment Rvp. The speed increment "Vf" and the speed increment 1■9 are input, and the speed control factor calculation means 86 calculates the moving speed RvH.The control signal generating means 87 inputs the inverse Jacobian matrix J-1 and the moving speed RvH, and calculates the moving speed RvH from these. Calculate joint velocity θ.

The obtained joint velocity is output to the robot main body 20 via the compensator 88 and the DA converter.

In this way, in the teaching control system according to the embodiment, teaching is performed by grasping the connection point 1a between the hand section 3 and the force sensor section 4, and when reproducing the work, it is decided to reproduce it as an operation of the hand section 3. , the load on the hand section can be reduced when teaching, and there is no need to apply excessive load to the fingers of the hunt section 3.
Furthermore, teaching can be performed even when the hand section 3 is gripping the object.

〔Effect of the invention 〕

As described above, in the present invention, when teaching is performed by holding the connection point 1a, the force applied by the operator is detected by the force detection means 12 with the center point E of the connection point 1a as the point of force application, and the force applied by the operator is detected at the detected point E. When the force and its direction are input to the controller side and the robot 10 executes the work, the point detected by the force detection means 12 by the force processing means 14 is used as a point on the teaching trajectory drawn by the point of action H of the hand section 3. The teaching force F is processed, a transformation matrix for the coordinate transformation is calculated by the Jacobian matrix calculation means 15, and the control amount information generation means 17 inputting these results calculates the homogeneous transformation element of the three-dimensional coordinate transformation as a parameter. The equation is processed to obtain drive control information for the manipulator 1 necessary for the work operation in the hand part 3, and based on the result, the manipulator 1 is made to perform an action corrected for the force and direction of the force in the hand part 3. This eliminates the need for the teaching worker to be particularly conscious of the discrepancy between the gripping position of the worker and the position of the reproducing operation when teaching, making it easy to teach, and making it possible to avoid excessive pressure on the fingers of the hand unit 3 when teaching. This eliminates the need to apply heavy loads.
The load on the hand section can be reduced, and teaching can be performed even when the hand section 3 is gripping an object.

[Brief explanation of drawings]

FIG. 1 is a configuration explanatory diagram showing a robot teaching control system according to the present invention, FIG. 2 is an explanatory diagram showing robot direct teaching according to the present invention together with a conventional method, and FIG. 3 is an apparatus according to an embodiment. FIG. 4 is a block diagram showing the configuration; FIG. 4 is a configuration explanatory diagram showing the teaching control method of the speed control multi-articulation robot according to the embodiment; FIG. 5 is a perspective explanatory diagram showing the coordinates of the hand part in the embodiment; FIG. 6 is an explanatory diagram showing the interrelationship of each coordinate system in the embodiment, FIG. 7 is a flowchart showing the parameter replacement procedure by the host computer in the embodiment, and FIG. A configuration explanatory diagram showing another aspect of the teaching control method of the articulating robot, 1... Manipulator 1a... Connection point 3... Hand part 4... Force sensor part 10... Robot 12...・Force detection means 13...Drive means 13a...Position detection means 14...Force processing means 15...Jacobian matrix calculation means 16...Position processing means 17...Controlled amount information generation means λ ＋−ψ 口4 Rōroku (V

Claims (1)

[Claims] A robot (10) having a hand part (3) and a connection point (1a) between the hand part (3) and a force sensor part (4) at the tip of the manipulator (1), The force sensor section (4) is provided with a manipulator (
force detection means (12) for detecting the force applied to the manipulator (1); and a drive means (13) for inputting the force detected by the force detection means (12) and for driving the manipulator (1).
The displacement caused by the operation of the manipulator (1) is inputted from the position detecting means (13a) provided at the position detecting means (13a), and the parameters at each position on the teaching trajectory are inputted, and the reference point at each position on the teaching trajectory is inputted. A force processing means (14) that outputs force in a coordinate system; and a transformation for obtaining control amount information from a transformation matrix input from the force processing means (14) and a Jacobian matrix at the position of the force detection means (12). Jacobian matrix calculation means (15) for calculating a matrix, and the position detection means (13a) based on the displacement inputted from the hand section (3) or the connection point (1a).
a position processing means (16) for outputting the current position in the position processing means (16); a current position input from the position processing means (16); and a transformation matrix input from the Jacobian matrix calculation means (15);
From the force of the reference point coordinate system input from the force processing means (14), the hand part (3) or the connection point (1a
control amount information generating means (
17), wherein when teaching the work of the manipulator (1), the connecting part (1a) is manually operated, and when reproducing the work, the robot is regenerated as the movement of the hand section (3). control method for teaching.