JP3433465B2 - Robot learning control method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a robot learning control method suitable for use in, for example, an industrial robot performing a repetitive reproduction operation. 2. Description of the Related Art As is well known, various techniques for repeatedly reproducing a robot have been developed. In this type of technology, for example, a learning control method described in the following document (a) is known. (A) S.I. Arimoto: Learning C
ontrol Theory for Robotic
Motion, International Jou
rnal of Adaptive Control
and Signal Processing, 4-
6, 546/564 (1990) In the learning control method described in this document (a), a negative feedback loop of a position error and a speed error is individually formed for each degree of freedom of the robot. Then, the robot performs a reproducing operation in accordance with a target trajectory of a joint corresponding to each degree of freedom prepared in advance. Then, the speed error between the target trajectory at this time and the actual trajectory generated during the reproducing operation is measured, and this speed error is multiplied by a constant (hereinafter, referred to as a learning gain). Further, the result of the multiplication is added to an input value (hereinafter, referred to as a feedforward value) given to the driving element of the robot during the previous reproduction operation,
The accuracy of the reproduction operation is improved by providing the addition result as a new feedforward value to the drive element of the robot at the next reproduction operation. According to this learning control method, the actual trajectory of the robot is successively corrected while the reproducing operation is repeated several times,
It converges to the vicinity of the target trajectory prepared in advance.
Therefore, dynamic effects such as centrifugal force and Coriolis force seen in a robot having multiple degrees of freedom are finally corrected, and high-speed and high-accuracy trajectory control can be performed. Also, theoretically, the larger the value of the learning gain, the smaller the actual trajectory converges to the vicinity of the target trajectory with a smaller number of reproduction operations. However, in the above-described conventional learning control method, if the learning gain by which the speed error is multiplied is too large, the measurement error or the like that occurs at the time of speed detection at each joint causes an influence. The real orbit oscillates toward the end of the orbit. This is introduced, for example, by the following document (b). (B) Nanjo et al .: Effect of forgetting factor on robot trajectory control using learning, Proceedings of the 9th Annual Conference of the Robotics Society of Japan. 1, 391/392 (1991) FIG. 3 shows a conventional learning control method.
An example of an actual trajectory when a reproduction operation is performed 20 times when the learning gain is excessively applied when applied to a robot having a degree of freedom is shown. As shown in FIG.
The actual trajectory A2, which almost coincides with the target trajectory A1, vibrates toward the end of the trajectory. That is,
The conventional learning control method has a practical limitation on the magnitude of the learning gain, and therefore, a large number of reproduction operations must be performed to converge the actual trajectory to the vicinity of the target trajectory. There was a disadvantage. The present invention has been made under such a background, and an object of the present invention is to provide a robot learning control method capable of improving the learning efficiency of a robot having a learning function. According to the present invention, in order to solve the above-described problems, when a robot repeatedly performs a reproducing operation, a target trajectory which is a target of the reproducing operation and a target trajectory at the time of the reproducing operation are described. A learning control method for correcting an input value supplied to a driving unit of a robot based on an error occurring between the actual trajectory of the robot and a position error, a velocity error, and an acceleration error between the target trajectory and the real trajectory. A second step of multiplying the error measured in the first step by a coefficient (learning gain) that decreases as time elapses from the start of the reproducing operation. A third step of adding the multiplication result of the second step and an input value (feed forward value) supplied to the drive unit at the time of the previous reproduction operation, and the third step Is supplied to the drive unit as an input value (feed forward value) in the next reproduction operation.
And a step. According to the present invention, in the first step, any one of a position error, a velocity error, and an acceleration error between the target trajectory and the actual trajectory is measured, and in the second step, the error is measured. In the third step, the error measured in one step is multiplied by a coefficient (learning gain) that decreases with the lapse of time. In the third step, the multiplication result of the second step is multiplied by the input value ( Feed forward value), and in the fourth step, the third
The result of the step is supplied to the drive unit as an input value (feedforward value) in the next reproduction operation. Thus, if the feedforward value to the drive unit of the robot is repeatedly corrected for each reproduction operation, the vibration of the actual trajectory that occurs as the trajectory approaches the end is suppressed. Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a control circuit used in one embodiment of the present invention. In this figure, only the control system for one degree of freedom of the electric motor drive type robot is shown, and the other degrees of freedom have the same configuration and are not shown. In FIG. 1, reference numeral 4 denotes an arithmetic processing unit for performing various digital operations, which comprises a CPU (Central Processing Unit), memories 1 to 3, and the like (not shown). The arithmetic processing unit 4 includes an input / output interface circuit such as a D / A (digital / analog) conversion element, but is not shown here for simplicity. Each of the memories 1 to 3 has a storage capacity enough to store at least the number of data obtained by dividing the time T required to move the target trajectory by the control period T _S.
And the trajectory data θ of the target trajectory is stored in the memory 2.
_ref (t) is the target position θ of the motor 9 for each control cycle time T _S
ref (T _S i) (where, i is an integer) is previously stored as. The storage contents of the memories 1 and 3 and their uses will be described later. The arithmetic processing unit 4 is connected to a motor 9 for driving the robot via a comparator 5, a multiplier 6, a comparator 7, and an amplifier 8. Further, the motor 9 is connected to the comparator 7 via a rotation speed detector 11 and a multiplier 10 and is also connected to the comparator 5 via a rotation position detector 12. The comparator 5 receives the control cycle time T _S from the memory 2.
The difference between the target trajectory data θ _ref (t) supplied for each time and the actual trajectory θ (t) measured by the position detector 12 is calculated,
Output this result. The multiplier 6 multiplies the output of the comparator 5 and the positive constant k _1, and outputs the result. This multiplier 6
Is supplied to an amplifier 8 via a comparator 7. On the other hand, the speed detector 11 measures the actual speed dθ / dt of the motor 9 (hereinafter referred to as θ ^* (t)) and outputs the measurement result. The multiplier 10 multiplies the actual speed theta ^* measured by the speed detector 11 (t) and a positive constant k _2, and outputs the result. The output of the multiplier 10 is supplied to the amplifier 8 via the comparator 7 after the sign is inverted. Further, the amplifier 8 amplifies the output of the comparator 7 and outputs this to the motor 9. The above is the feedback control system in this control circuit. According to such a configuration, the motor 9 is driven in a direction in which the actual trajectory θ (t) has an error “0” with respect to the target trajectory θ _ref (t). At this time, in order to ensure the stability of the feedback control system, the values of the positive constants k ₁ and k ₂ cannot be extremely increased practically. When the robot is operated at high speed, the motor 9 cannot completely follow the target trajectory θ _ref (t) due to the influence of the centrifugal force and Coriolis force seen in a robot having many degrees of freedom. Therefore, in order to correct an error based on such a cause, feedforward control described below is performed in addition to the above feedback control. Next, the feedforward control operation based on the operation of the CPU will be described with reference to the flowchart shown in FIG. In the figure, first, in step p1, "0" data u _k is stored in the memory 1 is cleared. Here, the data u _k is a feedforward value given to the driving element of the robot at the time of the k-th reproduction operation. In this case, since it is the time of the first reproduction operation, the feed forward value u ₁ becomes “0”. Next, the process proceeds to step p2, in the feedback control system shown in FIG. 1, is supplied to the amplifier 8 the feedforward value u ₁ in the memory 1 for each control cycle T _S,
The first reproduction operation is performed. At this time, the velocity θ ^{* of the} entire real trajectory measured by the velocity detector 11 for each control cycle T _S
The (T _S i) stored in the memory 3. Then, when the first reproduction operation is completed, the feedforward value in the memory 1 is corrected by the following steps p3 to p5. At step p3, a target speed θ ^* _ref (t) is calculated based on the target trajectory θ _ref (t) stored in the memory 2. That is, as shown in the following equation (1), the time t
= T _S velocity at ^{_{i θ * ref (T S i}} ) can be calculated by dividing the difference between the target position in the control cycle before and after the control period T _S. Here, i is an integer representing the order of the control cycle. ^{_{θ * ref (T S i)}} = {θ ref (T S (i + 1)) - θ ref (T S i)} / T S ......... (1) [0020] Next, the process proceeds to step p4 The difference between the speed θ ^* _ref (t) calculated in step p3 and the actual speed θ ^* (t) in the previous reproduction operation stored in the memory 3 is calculated for each control cycle T _S , The result is multiplied by the learning gain φ. Here, the learning gain φ is determined by a function represented by the following equation (2), for example, so that the learning gain φ becomes smaller as approaching the end of the trajectory. Here, φ ₀ and λ are positive constants, and e is a natural logarithm. (Equation 1) Next, at step p5, the value obtained by multiplying the learning gain φ and the feedforward value u in the memory 1 are calculated.
₁ (t) is added to the new feedforward value u ₂
(t) is stored in the memory 1. And step p6
Then, it is determined whether to continue the learning operation. That is, if the number of times of the reproduction operation has not reached the preset number of learning times or the actual trajectory has not reached the preset desired accuracy, the determination result here is “Yes”, It returns to the above-mentioned step p2.
Thereby, the learning operation accompanying the second reproduction operation is performed. Next, in the second learning operation, the feedforward value u ₂ (t) newly stored in the memory 1 is supplied to the amplifier 8 in step p5. Thereafter, while the determination result of step p6 is "Yes", steps p3 to p5 are repeated for each reproducing operation, and the feedforward value in the memory 1 is corrected each time. The equation for calculating the feedforward value at this time is represented by the following equation (3). (Equation 2) In this way, when the number of times of the reproducing operation reaches the preset number of times of learning, or when the actual trajectory reaches the desired accuracy, the judgment result of step p6 is "N
o ", and the learning operation ends. As described above, according to the present embodiment, the learning gain for multiplying the difference between the target trajectory and the actual trajectory is reduced toward the end of the trajectory, thereby suppressing the vibration generated as the trajectory ends. can do.
As a result, the learning gain can be set larger than in the conventional method, and the real trajectory can be made to converge to the vicinity of the target trajectory with less repetitive reproduction operations. In this embodiment, the learning gain φ (t) in the above equation (3) is set by the function represented by the above equation (2), but a function value that decreases with the passage of time can be obtained. Other functions can also be used. Further, in this embodiment, it has been to calculate the target speed θ ^* _ref (t) based on the theta _ref (t) stored in the memory 2, alternatively, pre-target velocity theta ^* _ref (t) may be stored in another memory. Also, without storing the actual speed θ ^* (t) during the reproducing operation in the memory 3, the feedforward value for each control cycle is directly calculated by the above equation (3), and the calculated value u _{k + 1} (t ) May be stored directly in the memory 1. In this embodiment, the case where a computer (arithmetic processing unit 4 composed of a CPU or the like) is used as a control circuit is described as an example. However, this may be replaced with various other known control elements. It is possible. This embodiment is directed to a playback type robot in which each joint of the robot is driven by an electric motor, and a negative feedback control system for a position error and a speed error is individually configured for the electric motor. Although the present invention has been described, the present invention is also applicable to a robot having a hydraulic cylinder as a drive source, and other machines having various servo mechanisms. Further, in this embodiment, only the velocity error generated between the target trajectory and the actual trajectory of the robot is used for calculating the feedforward value. Learning can also be performed by adding the multiplied value to the feedforward value at the time of the previous reproduction operation. Further, the present invention can be used in combination with the learning control method described in the aforementioned reference (a). That is, in this document (a), in order to reduce the influence of measurement errors and the like, as shown in the following equation (4), a positive constant (1-a) suitable for the previous feedforward value u _k (t) is used. Has been proposed, and can be used together with the present invention. u _{k + 1} (t) = φ (θ ^* _ref (t) −θ ^* (t)) + (1−a) u _k (t) (4) As described above, according to the present invention, the vibration of the real trajectory which occurs as the trajectory approaches the end is suppressed, so that the learning gain can be set to be larger than in the past, and less repetitive reproduction is performed. The operation allows the actual trajectory to converge near the target trajectory. As a result, the learning efficiency of the robot can be improved.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a configuration of a control circuit used in one embodiment of the present invention. FIG. 2 is a flowchart showing a feedforward control operation based on an operation of a CPU in the embodiment. FIG. 3 is a diagram illustrating an example of an actual trajectory when a reproduction operation is performed 20 times when a conventional learning control method is applied to a robot having three degrees of freedom and a learning gain is excessively increased. [Description of Signs] 1 to 3 Memory 4 Arithmetic processing unit 5, 7 Comparator 6, 10 Multiplier 8 Amplifier 9 Motor 11 Rotation speed detector 12 Rotational position detector A1 Target trajectory A2 Actual trajectory

Continued on the front page (58) Fields surveyed (Int. Cl. ⁷ , DB name) G05B 19/18-19/46 B25J 3/00-3/04 B25J 9/10-9/22 B25J 13/00-13 / 08 B25J 19/02-19/06 G05B 13/02

Claims (1)

(57) [Claims 1] When a robot repeatedly performs a reproducing operation, an error generated between a target trajectory which is a target of the reproducing operation and an actual trajectory of the robot at the time of the reproducing operation is reduced. Based on
A learning control method for correcting an input value supplied to a drive unit of the robot, wherein a first error is measured among a position error, a speed error, and an acceleration error between the target trajectory and the actual trajectory.
Step, the error measured in the first step, and the reproduction operation start
A second step of multiplying by a coefficient that decreases as time elapses from time, a third step of adding the multiplication result of the second step and an input value supplied to the drive unit during a previous reproduction operation, And a fourth step of supplying the result of the addition in the third step to the drive unit as an input value in the next reproduction operation.