Machine learning device that adjusts controller gain in a servo control apparatus
A machine learning device that performs reinforcement learning with respect to a servo control apparatus that controls target device having a motor, including: outputting action information including adjustment information of coefficients of a transfer function of a controller gain to a controller included in the servo control apparatus; acquiring, from the servo control apparatus, state information including a deviation between an actual operation of the target device and a command input to the controller, a phase of the motor, and the coefficients of the transfer function of the controller gain when the controller operates the target device based on the action information; outputting a value of a reward in the reinforcement learning based on the deviation included in the state information; and updating an action-value function based on the value of the reward, the state information, and the action information.
1. A machine learning device that performs reinforcement learning with respect to a servo control apparatus that controls an operation of a control target device having a motor, the machine learning device comprising:
a memory configured to store a program: and
a processor configured to execute the program and control the machine learning device to:
output action information including adjustment information of coefficients of a transfer function of a controller gain to a controller included in the servo control apparatus;
acquire, from the servo control apparatus, state information including a deviation between an actual operation of the control target device and a command input to the controller, a phase of the motor, and the coefficients of the transfer function of the controller gain when the controller operates the control target device on the basis of the action information;
output a value of a reward in the reinforcement learning on the basis of the deviation included in the state information; and
update an action-value function on the basis of the value of the reward, the state information, and the action information so that the action information is updated and the coefficients of the transfer function of the controller gain are adjusted according to the value of the reward, the action information, and the state information including the phase of the motor so that uneven rotation of the motor is improved.
2. The machine learning device according to claim 1 , wherein
the servo control apparatus is a servo control apparatus that performs feedback control for correcting the command input to the controller, and
the processor is further configured to control the machine learning device to acquire a difference between the command input to the controller and a feedback value of the feedback control as the deviation.
3. The machine learning device according to claim 1 , wherein
the controller is a combination of controllers that perform position control, speed control, and current control, and
the machine learning device selects the controller that performs current control, the controller that performs speed control, and the controller that performs position control in that order as a reinforcement learning target when the machine learning device performs the reinforcement learning by selecting any one of the controllers as a target and then performs the reinforcement learning by selecting another controller as a target.
4. The machine learning device according to claim 1 , wherein
the phase of the motor is calculated on the basis of a position command for controlling the operation of the control target device.
5. The machine learning device according to claim 1 , wherein
the transfer function of the controller gain includes a phase of the motor as a variable.
6. A servo control system including the machine learning device according to claim 1 and the servo control apparatus, wherein
the servo control apparatus includes
a memory configured to store a program: and
a processor configured to execute the program and control the servo control apparatus to:
calculate a phase of the motor on the basis of a position command for controlling an operation of the control target device and output the calculated phase of the motor to the machine learning device and the controller.
7. A machine learning method of a machine learning device that performs reinforcement learning with respect to a servo control apparatus that controls an operation of a control target device having a motor, the machine learning method comprising:
an action information output step of outputting action information including adjustment information of coefficients of a transfer function of a controller gain to a controller included in the servo control apparatus;
a state information acquisition step of acquiring, from the servo control apparatus, state information including a deviation between an actual operation of the control target device and a command input to the controller, a phase of the motor, and the coefficients of the transfer function of the controller gain when the controller operates the control target device on the basis of the action information;
a reward output step of outputting a value of a reward in the reinforcement learning on the basis of the deviation included in the state information; and
a value function updating step of updating an action-value function on the basis of the value of the reward, the state information, and the action information so that the action information is updated and the coefficients of the transfer function of the controller gain are adjusted according to the value of the reward, the action information, and the state information including the phase of the motor so that uneven rotation of the motor is improved.