JPH0546207A - Learning control method - Google Patents

Abstract

PURPOSE:To improve the follow-up precision by making a learning starting spot into not the starting spot of a target command but the spot where deviation is zero or the spot near it. CONSTITUTION:In a learning control system to repeat the trial so as to follow up the controlled to a target command to repeat the same pattern, a deviation deciding part is provided, and it decides whether or not the difference between the position of a motor and the target position command becomes zero. When the difference does not become zero, it is decided that the spot near the zero is zero. When the decision is performed, a learning starting command is generated to a learning controller and the learning is started. Even when the special deviation deciding part is not provided, realization is possible by changing the software in the learning controller.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of controlling a machine tool, a robot or the like which repeatedly operates.

[0002]

2. Description of the Related Art As a method of designing a learning control system for a repetitive target value, there is a method proposed by the applicant in Japanese Patent Laid-Open No. 1-237701. This method predicts future deviations based on past deviations, corrects the control input so that the value is minimized, and realizes highly accurate tracking operation.When applied to motor position tracking control. Is shown in FIG. In addition, the present applicant has proposed in Japanese Patent Application No. 2-196940, in which the difference between the position of the motor detected by the position detector of the motor and the target position command is used as the speed command, and based on the speed command. FIG. 4 shows a device that realizes a highly accurate follow-up operation by correcting the speed command so that the value becomes minimum.

[0003]

However, when the learning calculation is performed using the above method, conventionally, the learning start point is the start point of the position command as shown in FIG. 5, and the learning calculation is repeated even if the learning calculation is repeated. There is a whisker-like deviation in the vicinity of each of the starting points, and it takes a long time to converge to a deviation of a predetermined size, and in some cases, it does not converge. Therefore, an object of the present invention is to provide a method of reducing the deviation near the learning start point and allowing the deviation of a predetermined size to converge more quickly.

[0004]

In order to solve the above problems, in the present invention, in a learning control system in which trials are repeated, a learning start point is set as a target so that the controlled object follows a target command in which the same pattern is repeated. The feature is that the learning start point is not the start point of the command but the deviation is zero, or the point near zero is set as the learning start point.

[0005]

With the above means, the learning start point converges to a predetermined deviation faster than when the learning start point is the target command start point.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A specific embodiment of the present invention will be described below with reference to an example applied to a positioning control system using a servo motor. 1 and 2 are specific embodiments of the present invention, respectively.
This is an example applied to a control system of the type shown in FIGS. 3 and 4 showing a conventional example. In order to carry out the method of the present invention, a deviation determination unit is provided as shown in the figure, and it is determined whether or not the difference between the motor position and the target position command becomes zero (point T in FIG. 5 shows zero deviation). Point). If it does not become zero, points near zero are determined to be zero. If the judgment is made,
A learning start command is issued to the learning controller to start learning. As described above, the present invention can be applied to any type of control system including a learning controller. Further, it can be realized by changing the software inside the learning controller without providing a special deviation determining unit. As an example when applied to the learning control system by correction of the speed command shown in FIG. 1, the transfer function from the corrected speed command u to the speed v is 1 /
(S ² + 2ζωs + ω ² ) and the damping coefficient ζ is 1,
The loop gain Kp is 100 (rad / sec) and the angular frequency ω is 300 (ra
FIG. 6 shows an operation explanatory diagram of the case of d / sec). Comparing this with FIG. 7, which is an operation explanatory diagram when learning is started at the conventional position command start point, it can be seen that the deviation is significantly reduced by changing the learning start point. It should be noted that in FIGS. 6 and 7, the deviation becomes small after the learning is started, so that the deviation is enlarged by 20 times in the middle for easy viewing.

[0007]

As described above, according to the present invention, in the learning control system in which trials are repeated so that the controlled object follows the target command in which the same pattern is repeated, the learning start point is the starting point of the target command. Instead, by setting the point at which the deviation becomes zero or the vicinity thereof as the learning start point, the deviation can be converged to the target deviation more quickly, and the following accuracy is improved.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing an embodiment of the present invention.

FIG. 3 is a diagram showing a conventional example.

FIG. 4 is a diagram showing a conventional example.

FIG. 5 is an operation explanatory diagram of a conventional example.

FIG. 6 is an operation explanatory diagram of the present invention.

FIG. 7 is an operation explanatory diagram of a conventional example.

Claims (2)

[Claims] 1. A learning control system in which trials are repeated so that a controlled object follows a target command that repeats the same pattern, and a learning start point is a learning start point at which a deviation becomes zero. Learning control method. 2. In a learning control system in which trials are repeated so that a controlled object follows a target command that repeats the same pattern, if the deviation does not become zero at the learning start point,
A learning control method characterized in that a point where the difference is minimized is set as a learning start point.