Research On High-precision Servo Control Technology Of ATP System In Space Optical Communication

The field of optical communication is currently focused on the research of optical space communication technology,which is known for its speed,capacity,and confidentiality.The Acquisition,Tracking,and Pointing(ATP)system is a crucial condition and vital technology that enables optical space communication.Within the ATP system,signal light is preliminarily captured and targeted through rough tracking to establish the foundation for communication.The reduction of overshoot and improvement of control accuracy in the servo execution mechanism of rough tracking are important research topics in the field of optical space communication.The aim of this thesis is to improve the control accuracy of the permanent magnet synchronous motor(PMSM)within the ATP system of optical space communication.The thesis is conducted taking into account the input constraints and requirements of the iindicators with the objective of decreasing the overshoot of the permanent magnet synchronous servo motor and enhancing its control accuracy.This thesis presents the structure and working principle of PMSM,selects vector control as the control strategy,and accomplishes the three-loop control of PMSM and the implementation of the Space Vector Pulse Width Modulation Algorithm(SVPWM)as well.Furthermore,the PMSM control system model based on parameter identification and PID is constructed on the Simulink platform.Simulation and verification indicate that the parameter identification PID control system fails to meet the development requirements due to its high overshoot and inadequate control accuracy.To address the inadequate control accuracy of the PID algorithm,this thesis conducts an extensive analysis of the Deep Deterministic Policy Gradient(DDPG)algorithm.This thesis presents the algorithm development of DDPG,designs the current loop and speed loop for the control strategy,and trains two reinforcement learning agents concurrently.Different segmented reward functions are selected for the two agents to improve training results.Using this algorithm,the thesis designs and simulates the permanent magnet synchronous motor servo control system on the Simulink platform,and experimental results demonstrate that the DDPG algorithm offers higher control accuracy.Using the experimental platform of optical space communication,this thesis designs the corresponding motor control circuit,physically implements it via DDPG training performed offline,and compares this technique with the PID control method.Experimentation confirms that the control system based on DDPG algorithm efficiently improves control accuracy and meets the development requirements.