JP2000176868A - Robot control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically learn and master optimum solutions such as a parameter or a moving route by moving a work in a driven action mode following to a specified route based on the respective assumed values of an optimized total weight and relative position. SOLUTION: A robot 10 begins movement from the initial position, and grasps a work W with a hand 17. A position of the total weight (m) of the work W and the hand 17 is expressed by a vector (r). However, the beginning point of the vector (r) is the reference point O on a wrist part 14, and the terminal point is the center of gravity of the total weight (m). Because (m), (r) are assumed values decided from the model on a drawing, they do not necessarily coincided with actual values. One points on the one extreme end of the work W grasped with the hand 17 is set as Wa, and the position of Wa at completing grasping action is set as P2. The robot 10 carries the work W while grasping it so that Wa reaches to the beginning point P0 on the specified route C0 defined on a space.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for a robot that operates in response to an external force acting on a tool.

[0002]

2. Description of the Related Art Conventionally, when an external force is applied to a tool, a current limit is provided to a servomotor that drives a robot shaft so that the shaft has a softness and the robot tool is moved in a direction to absorb the external force. Was moved so that the robot could follow the external force. Then, parameters (arm length, mass, etc.) for calculating the gravitational torque for each posture are given, the gravitational torque is calculated using the angles of the respective axes, and control is performed so that the gravitational torque is balanced.

[0003]

The various parameters (arm length, mass, etc.) used in the calculation of the gravitational torque are often numerical values obtained from drawings such as design drawings of robots and works. In the calculation of the gravitational torque using these numerical values (parameters), an accurate result cannot be obtained in many cases. Therefore, in a parameter setting operation in a robot control system that requires smoother control, it is necessary to obtain a proper numerical value by repeating parameter resetting by trial and error, and the operation efficiency is poor. Further, when the work is moved along a movement path of a complicated shape accompanied by an external force, it may be difficult to accurately teach the movement path of the work to the robot.

[0004] The present invention has been made to solve the above-mentioned problems, and has as its object the function of the driven robot to automatically learn and acquire the optimal solution such as the above-mentioned parameters and moving paths. It is to provide.

[0005]

In order to solve the above-mentioned problems, the following means are effective. That is, the first means includes a plurality of axes driven by a servomotor,
A robot controller having a driven operation mode that is driven in response to an external force acting on the outside is provided with a hypothetical value of the total weight of the weight of the robot hand and the weight of the work, and the center of gravity of the total weight and the robot center. Assumption value setting means for setting an assumption value of a relative position from the wrist or the arm, and recording a current value of the servomotor and a rotation angle of the servomotor while moving the work in the driven operation mode according to a predetermined path. Movement progress recording means, based on a current value or a rotation angle recorded by the movement progress recording means, an assumption value optimization means for correcting the above assumption values of the total weight and the relative position, and an assumption value optimization means Automatic driving means for moving the work according to a predetermined path in the driven operation mode based on the assumed values of the total weight and the relative position optimized by the above.

The second means includes a robot control device having a plurality of axes driven by a servomotor and having a driven operation mode that is driven in response to an external force applied from the outside. Assumed value of the total weight of the weight of the robot and the weight of the work, assumed value setting means for setting the center of gravity of the total weight and the assumed value of the relative position from the wrist or arm of the robot, and the following operation according to a predetermined path. A moving path recording means for recording the current value of the servomotor and the rotation angle of the servomotor while moving the work in the mode; and a predetermined path based on the current value or the rotation angle recorded by the moving progress recording means. It is provided with a route optimizing means for correcting, and an automatic driving means for moving a work in a driven operation mode in accordance with a predetermined path optimized by the route optimizing means.

Further, the third means comprises the first and second means.
Are used in combination.

Further, the fourth means includes the first to third means.
In the assumption value optimizing means or the path optimizing means in any one of the means, these corrections are performed by rotating the servo motor at each point on a predetermined path calculated from the current value recorded by the movement progress recording means. This is performed by minimizing the difference between the torque value, the assumed value of the total weight, the assumed value of the relative position, and the rotational torque value of the servo motor for each point on the predetermined path calculated from the predetermined path.

Further, the fifth means includes the first to fourth means.
In the automatic driving means in any one of the means,
When the current value or the rotation angle recorded by the movement progress recording means exceeds the upper limit value or the lower limit value for the current value or the rotation angle defined for each point on the predetermined path, the above-described hypothetical value optimization means or Restarting the route optimization means. With the above means, the above-mentioned problem can be solved.

[0010]

According to the first means, the current value of the servo motor and the rotation angle of the servo motor recorded by the robot controller while the workpiece is moved in the driven operation mode according to the predetermined path. Are corrected based on the above assumptions of the total weight and the relative position. Therefore, if this correction is repeated a necessary number of times, the above-described optimization of each assumed value of the total weight and the relative position can be automatically performed. Therefore, if the work is moved along the predetermined path in the driven operation mode based on the optimized total weight and the assumed values of the relative position, desired smooth control can be realized. Since these optimizations are automatically performed by the robot, the operator only needs to set the above assumed values of the total weight and the relative position by the assumed value setting means, thereby improving the efficiency of the teaching operation. I do.

According to the second means, the predetermined path can be corrected based on the recorded current value or the rotation angle. Therefore, if this correction is repeated a required number of times, the movement path of the work can be optimized. Can be Therefore, if the workpiece is moved according to the optimized movement path (predetermined path), desired smooth control can be realized.

Further, according to the third means, it is possible to simultaneously obtain the functions and effects of the first and second means. Further, according to the fourth means, it is possible to more specifically, mechanically, and efficiently realize the above-described correction using information that can be detected by the robot control device.

According to the fifth means, the assumption value optimizing means or the route optimizing means is started at any time when it is required, so that a large number of works are repeatedly operated by the robot. In this case, it is more effective to realize desired smooth control, particularly when there is a large variation in the weight and length of the work.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described based on specific embodiments. (First Embodiment) FIG. 1 is a schematic configuration diagram of a robot according to the first and second embodiments. In the six-axis vertical articulated robot 10, the rotator 11 is supported on a fixed base 27 so as to be rotatable in a horizontal plane about a vertical axis (one axis) X <b> 1, and the first arm 12 is placed horizontally on the rotator 11. It is supported so as to be swingable about an axis (biaxial) X2. The first arm 12 has a second arm 13 having a horizontal axis (3
The shaft is supported so as to be swingable about X3.

On the rotator 11, a horizontal axis X
The first link 18 is supported so as to be swingable about the center 2, and the first link 18 and the end of the second arm 13 are connected by the second link 19. When the first link 18 is driven, the second arm 13 swings. Second arm 1
3 has a wrist 1 having four axes, five axes and six axes
4 is rotatably supported. A hand (tool) 17 having a pair of gripping claws 17a (see FIG. 4) is provided on a flange at the tip of the wrist portion 14.

Next, the configuration of the control device of the robot 10 will be described with reference to FIG. The CPU 80 is connected with a memory 81, servo units 91 to 96 corresponding to each axis, and a teaching box 70. Each of the servo units 91 to 96 includes a servo CPU and a memory, and outputs a command rotation angle signal θ output from the CPU 80.
1 to θ6, gravity torque values Gf1 to Gf6, inertia value JL1
The servo motors M1 to M6 used for driving the 1-axis to 6-axis are controlled based on JL6 to JL6.

The encoders E1 to E6 connected to the servomotors M1 to M6 detect and output the rotation angles α1 to α6 of the servomotors M1 to M6. The current detectors 61 to 66 connected to the servomotors M1 to M6 detect and output the currents β1 to β6 of the servomotors M1 to M6. The outputs α1 to α6 and the outputs β1 to β6 are
0 and input to the servo units 91 to 96,
It is used for the calculation of the gravitational torque value and the inertia value of each axis by 0 and the position feedback control and the speed feedback control by the servo units 91 to 96.

The memory 81 includes a program area for storing an operation program of the robot 10, a processing data area for storing data required for processing such as teaching points, and outputs α1 to α.
6, a control data area for storing outputs β1 to β6 and the like. The teaching box 70 is the robot 1
It is for inputting a teaching operation of 0 and an operation program, and includes a display 70a and a keyboard 70b for inputting an operation command to the robot 10, an operation program and the like.

Next, the details of the servo units 91 to 96 are shown in FIG. 3 and the processing contents of each part thereof will be described. The speed feed forward 407 is input to the command value after the position loop 401, and the gravity torque feed forward 405 and the acceleration torque feed forward 406 are input to the command value after the speed loop 402. Thereafter, a current limiting unit 403 is provided, and the current command value determined by the current limiting unit 403 is output to the amplifier unit 404. The current limiting unit 403 includes the gravitational torque feed forward 40
A current limit value is determined in consideration of a predetermined current value width with respect to the gravitational torque values Gf1 to Gf6 of the respective axes required to maintain the postures of the robot 10 which are output in step 5. These functions are achieved by digital processing.

The operation of the robot 10 based on the above configuration will be described. FIG. 4 is an explanatory diagram illustrating the movement route of the representative point Wa on the work according to the first embodiment. The robot 10 starts moving from an unillustrated initial position Px, and holds the workpiece W prepared on the mounting table 40 in FIG. The position of the total weight m of the work W and the hand 17 is represented by a vector r. However, the starting point of this vector r is the reference point O on the wrist 14, and the ending point of this vector r is the center of gravity of the total weight m. m and r are
Since it is a hypothetical value determined from a model on the drawing, it does not always match an actual numerical value. Hereinafter, one point on one end of the work W gripped by the hand 17 is defined as Wa.
The position of Wa at the time when this gripping operation is completed is P
Let it be 2. Then, the robot 10 transports while holding the workpiece W, as Wa comes to the start point P0 on a given path C ₀ that is defined in the space. The above operation is performed without providing the above current limitation.

The starting point P0 on C ₀ is a point set in front of the insertion opening BO of the object B into which the workpiece W is inserted, and extends from the starting point P0 to the end point P1 on the predetermined route C _0. driven to Wa is moved along this predetermined path C _0. However, the operation to be performed driven manner, when the workpiece W is subjected to external force, Wa is sometimes deviate from the predetermined path C _0. For example, in the figure 4, the tapered portion Wt of the workpiece W is a collision, to receive the drag force from the tapered portion Bt of the insertion port BO of the insert body B, and in fact, Wa is out of the predetermined path C _0, moving moving on path C _1. In the above-described driven operation, current limitation is performed on the servo motors M1 to M6.

The robot 10 moves to the position P1 where the insertion operation is completed.
After releasing the gripping of the work W at the position and releasing the work W, it returns to the initial position Px to complete one cycle. The position and orientation of the robot 10 at the positions of the points P0, P1, and P2 are stored in the memory 81 in advance as teaching point data. The teaching point data has been previously taught by the operator using the teaching box 70. Hereinafter, as shown in the following equation (1), the point P on a predetermined path C _0, when the control parameter s is increased from s0 to s1, it is assumed that moving from P0 to P1.

C ₀ = {P (s) | s0 ≦ s ≦ s1}, P (s0) = P0, P (s1) = P1. (1) The control parameter s is a function of the rotation angles α1 to α6. That is, assuming that the function is g, the control parameter s is given by the following equation (2).

S = g (α1, α2, α3, α4, α5, α6) (2)

FIG. 5 shows a servo motor M according to this embodiment.
4 is a graph showing torque fluctuations of FIG. The current β [A] of the motor M stored by the current detector can be described as a function of the control parameter s. At this time, assuming that this function is I, the current β of the motor M is expressed by the following equation (3).
Given by

Β = I (s) (3) The torque τ (s) shown by the solid line graph in FIG. 5 is obtained when the representative point Wa of the workpiece W moves on the movement path C _{1 in} FIG. Where τ (s) is given by the following equation (4).

Τ (s) = KI (s) (4) where K [V · s / rad] is a torque constant of the motor M.

On the other hand, the torque T shown by the broken line graph in FIG.
(S, m, r) is a torque value of the motor M calculated by a predetermined function T based on the model (m, r) on the drawing. Here, the cause of the inconsistency between τ (s) and T (s, m, r) is as follows. First, the model (m, r)
Is not an accurate model. Hereinafter, in the first embodiment, a means for optimizing the model (m, r) will be described.

FIG. 6 is a flowchart showing the assumed value optimizing process according to the first embodiment. Here, the “assumed value” refers to the model (m, r). In the present process, first, in step 705, the work W is moved according to the predetermined route C using the movement progress recording means for recording the outputs α1 to α6 and the outputs β1 to β6.
The value of I (s) at that time is obtained from each measured value. In step 710, the torque τ (s) is obtained from (4). In step 715, the assumed value saving area (μ, γ) of the model (m, r) is initialized to (0, 0), and the variable j is initialized to −1.

In step 720, the whole set U of model points (m _j , r _j ) near the assumed value (m ₀ , r ₀ ) first given by the worker is determined. For example, m _j
May be determined as in the following equation (5).

M _j = m ₀ {1+ (i / 1000) cos (iπ)} (i = 0, 1, 2,..., 50) (5) Candidate (m
_j , r _j ) is determined.

In step 725, all the elements of the entire set U
Element, ie, all model points (m_j, R_j) Goes to step 740
Check whether all items have been selected or not.
Go to step 730, otherwise go to step 735
Transfer. In step 735, the value of j is increased by one. S
In step 740, one unselected model point (m_j, R
_j) Is selected from the whole set U. In step 745
Is (m, r) = (m_j, R_j), The following equation (6)
Judge whether or not
To step 750, otherwise to step 725.
You.

[6] δf (s, m, r) ≡ {τ (s) -T (s, m, r)} 2 <δτ 2 (s0 ≦ ∀s ≦ s1) ... (6) However, .DELTA..tau function T And is a constant appropriately determined by the operator according to the specific application of the robot 10.

In step 750, (m, r) =
It is determined whether or not the following equation (7) holds for (m _j , r _j ). If so, the process proceeds to step 755; otherwise, the process proceeds to step 725.

Equation 7] δF (C, m, r) ≡∫ C δf (s, m, r) ds <ε ... (7) however, epsilon worker is appropriately determined depending on the specific application of the robot 10 Is a constant. The integration range of ds is
“S0 ≦ s ≦ s1”.

In step 755, the value of δF (C, m, r) obtained in step 750 is substituted for ε. In step 760, the value of (m _j , r _j ) is stored in the assumed value save area (μ, γ). In step 730, it is checked whether or not μ is determined. If μ> 0, the process proceeds to step 770; otherwise, the process proceeds to step 720. Step 7
At 70, the value of the assumed value saving area (μ, γ) is substituted for the model point (m, r) used in the automatic driving means.

When step 720 is executed a plurality of times, the following precautions are required in the second and subsequent executions. That is, the new overall set U to be determined again is determined so as not to include any elements that have been selected up to the previous time, ie, model points (m _j , r _j ) that have been selected up to the previous time. Must. When step 720 is performed a plurality of times, a new whole set U is determined again in the second and subsequent times. For example, in the second time, the range of possible values of m _j may be determined as in the following equation (8).

M _j = m ₀ {1+ (i / 1000) cos ((i−1) π)} (i = 1, 2,..., 100) (8)

The model points (m,
If r) is obtained, it becomes possible for the automatic driving means to perform a more accurate torque calculation using a more accurate model. In the above assumption value optimization processing, for simplicity, an optimal solution was obtained by a direct search method.
Such optimization processing is performed by a known descent method (descent me
For example, various optimization methods for calculating an optimal solution by a finite number of procedures, such as thod), may be used.

(Second Embodiment) In the first embodiment, τ (s)
And T (s, m, r) do not coincide with each other because the model (m, r) is not accurate, but the other cause is that they do not coincide with each other. It is difficult to accurately teach C to the robot 10 in advance. Hereinafter, in the second embodiment, a means for optimizing the movement route C will be described.

FIG. 7 shows a representative point W on the work W of this embodiment.
It is explanatory drawing which shows the moving route of a. As described above, when the shape of the work W or the inserted object B into which the work W is inserted is relatively complicated, the movement path C of the work W due to the following operation is accurately instructed to the robot 10 in advance. It is difficult. FIG. 8 is a flowchart illustrating the route optimization processing according to the second embodiment. Here, the “path” is the movement path C of the work W described above.

In this processing, first, at step 905
Initializes the variable j to 0. Next, step 9
In step 10, the movement path C for moving the representative point Wa of the work W is determined to be _Cj . Here, C ₀ is a predetermined path of the work W assumed by the operator as an initial value. Step 9
At 15, the variable j is increased by one. In step 920,
According to a predetermined path C _0, using the moving course recording means to record the actual movement path C _j and the current I (s) of the workpiece W.

In step 925, the torque τ (s) is obtained from (4). In step 930, (6),
It is determined whether or not (7) is satisfied. If so, the process proceeds to step 910; otherwise, the process proceeds to step 935. By repeating the above processing, the predetermined route C ₀ becomes the optimal moving route C. This is because the work W is controlled by the driven operation, so that when any drag occurs due to a collision or the like, the work W is pushed back to the more desirable movement path C. The above processing is repeatedly executed until (6) and (7) are satisfied, so that the route C is optimized so as to satisfy (6) and (7).

In step 935, the variable j is incremented by one. In step 940, it is determined whether or not the predetermined insertion process shown in FIG.
If j> N, the predetermined insertion processing ends, and if not, the processing moves to step 945. In step 945, the automatic operation means of said, at the same time moving the workpiece W in accordance with the movement path C optimized described above, using the moving course recording means of the actual movement path C _j and the current of the workpiece W Record I (s). In step 950, the torque τ (s) is obtained from (4). In step 955, it is determined whether or not (6) is satisfied. If so, the process proceeds to step 935; otherwise, the process proceeds to step 960. In step 960, the route C for moving the representative point Wa of the work W is determined to be _Cj .

When the path optimization processing is performed as described above, when the robot 10 repeatedly inserts a large number of works W, the movement path C is optimized again by the steps 950, 955, and 960 when necessary. Work W
This is effective in realizing a desired smooth control when there is a large variation in the weight of the insertion member B or the diameter of the insertion opening BO of the inserted object B.

In the above embodiment, the assumed value optimizing process and the route optimizing process are separately illustrated. However, these means can be used in an appropriate combination at the same time.

In the above-described embodiment, the case where the insertion process is performed has been described. However, the present invention is applicable to a case where the robot performs a follow-up operation with respect to a push-out operation such as a case of taking out a work from a die-cast machine. Also,
Also, the present invention can be applied to a case where a robot performs other general driven operations.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a robot according to first and second embodiments.

FIG. 2 is a hardware configuration diagram of a robot control device according to the first and second embodiments.

FIG. 3 is a functional block diagram of a servo unit according to the first and second embodiments.

FIG. 4 is an explanatory diagram showing a movement route of a representative point Wa on a work according to the first embodiment.

FIG. 5 is a graph showing torque fluctuation of a servo motor M according to the first embodiment.

FIG. 6 is a flowchart illustrating an assumption value optimization process according to the first embodiment.

FIG. 7 is an explanatory diagram showing a movement route of a representative point Wa on a work according to the second embodiment.

FIG. 8 is a flowchart illustrating a route optimization process according to the second embodiment.

[Explanation of symbols]

10 Robot 14 Wrist 17 Hand M Servomotor W Work B B Inserted object C _0, C into which work W is inserted C Predetermined path for moving work W P Point P0 on predetermined path C Start point P1 of predetermined path C: End point of predetermined path C s: Parameter indicating the position of point P on predetermined path C I (s): Current value of servo motor M at point P (s) m: Weight of hand 17 Total weight with the weight of the work W r ... Relative position of the center of gravity of the total weight m from the wrist 14 T ... Function indicating the torque value of the servo motor M at the point P (s) of the model (m, r)

Claims (5)

[Claims] 1. A control device for a robot having a plurality of axes driven by a servomotor and having a driven operation mode in which the robot operates in response to an external force applied from outside, the weight of a hand of the robot being reduced. An assumed value of a total weight of the weight of the work, an assumed value setting means for setting an assumed value of a center of gravity of the total weight and a relative position from the wrist or arm of the robot, and the driven operation mode according to a predetermined path. A movement progress recording means for recording a current value of the servo motor and a rotation angle of the servo motor during the movement of the workpiece, and based on the current value or the rotation angle recorded by the movement progress recording means. Assumed value optimizing means for correcting each of the assumed values of the total weight and the relative position; and the total weight and the relative position optimized by the assumed value optimizing means. Based on the assumed value, the robot control apparatus characterized by comprising an automatic driving means for moving the workpiece in accordance with the predetermined path by the driven mode of operation. 2. A control device for a robot having a plurality of axes driven by a servomotor and having a driven operation mode in which the robot operates in response to an external force applied from outside, wherein the weight of a hand of the robot is reduced. An assumed value of a total weight of the weight of the work, an assumed value setting means for setting an assumed value of a center of gravity of the total weight and a relative position from the wrist or arm of the robot, and the driven operation mode according to a predetermined path. A movement progress recording means for recording a current value of the servo motor and a rotation angle of the servo motor during the movement of the workpiece, and based on the current value or the rotation angle recorded by the movement progress recording means. A path optimization unit that corrects the predetermined path; and the work in the driven operation mode according to the predetermined path optimized by the path optimization unit. Robot control apparatus characterized by comprising an automatic driving means for moving. 3. The robot control device according to claim 1, wherein: 4. The rotational torque of the servo motor for each point on the predetermined path calculated from the current value recorded by the movement progress recording means, wherein the assumption value optimizing means or the path optimizing means is By minimizing the difference between the assumed value of the total weight, the assumed value of the relative position, and the rotational torque value of the servo motor for each point on the predetermined path calculated from the predetermined path, The robot controller according to any one of claims 1 to 3, wherein correction is performed. 5. The automatic driving means, wherein the current value or the rotation angle recorded by the movement progress recording means is an upper limit for the current value or the rotation angle defined for each point on the predetermined path. The robot controller according to any one of claims 1 to 4, wherein the assumption value optimizing unit or the route optimizing unit is restarted when a value or a lower limit value is exceeded.