A Reinforcement Learning Adaptive Control Method for Brushed DC Motor

Abstract

The invention discloses a brushed DC motor reinforcement learning self-adaptive control method, which belongs to the technical field of motors. The purpose of the present invention is to solve the problem of dead zone and crawling in the speed control process of brushed DC motors through online identification of cogging torque, nonlinear friction model parameters and unknown disturbance estimation and compensation strategy based on reinforcement learning through the parameter robust adaptive law. Non-linear control method of brushed DC motor speed and low speed instability problem. The steps of the invention are: establishing a mathematical model of the brushed DC motor; and a robust self-adaptive control method for the brushed DC motor. The present invention designs a feedforward plus feedback two-degree-of-freedom control structure based on differential flatness. Compared with the traditional double-loop PI control method, the feedforward control introduced by this method can act on the controlled object at the instant of reference input rather than after a deviation occurs, and a nonlinear compensation signal is introduced in the feedforward, which can suppress the impact of disturbance on the motor. The effect of low speed control improves the speed tracking accuracy.

Description

technical field [0001] The invention belongs to the technical field of motors. Background technique [0002] Brushed DC motor is an important industrial basic component, which has the advantages of large torque coefficient, strong overload capacity and high reliability, and is widely used in automobiles, robots, aerospace and other fields. With the rapid development of modern science and technology, especially the great progress of power electronics, digital control technology and modern control theory, favorable conditions have been created for the development of high-precision speed control of brushed DC motors. The high-precision speed control of brushed DC motors is subject to received more and more attention. The requirements for the control performance of brushed DC motors in many fields are constantly improving, and the difficulty of developing high-precision motor speed control methods has therefore become higher. [0003] Friction torque and cogging torque are two...

AI Technical Summary

Problems solved by technology

[0010] The purpose of the present invention is to solve the problem of dead zone and crawling in the speed control process of brushed DC motors through online identification of cogging torque, nonlinear friction model parameters and unknown disturbance estimation and compensation strategy based on reinforcement learning through the parameter robust adaptive law. Non-linear control method of brushed DC motor speed and low speed instability problem

Images