Active Disturbance Rejection Control Of Velocity Loop In Gyroscope Stabilized Platform Based On Genetic Algorithm

Gyro Stabilized Platform (GSP) is the core equipment in the inertial navigation, guidance and measurement system. It can effectively isolate the carrier disturbance and make the platform stable. As a consequence, it is able to ensure that the visual axis of the optical tracking devices (cameras, binoculars and weapons equipment, etc.) mounted on the platform can accurately point to a specific target. GSP is widely used in various military and civilian projects. As the demand of control accuracy in engineering applications increases, the traditional PID control technology is difficult to meet the growing performance requirements of GSP, so it’s necessary to design some new control technologies in GSP.Active Disturbance Rejection Control (ADRC) has got widespread attention by academics in recent years. It can estimate and compensate various disturbances in the controlled system in real-time, and does not depend on the exact model of the controlled system, which has strong robustness and is advantage of the classic PID control. A lot of researches on ADRC applied in engineering have been made to solve the problems which the traditional PID control is unable to solve.This dissertation studied the ADRC design of the two-axis and four-gimbal airborne electro-optical gyro-stabilized platform based on Genetic Algorithm. Specific work can be summarized as follows:(1) Based on the non-linear model of the velocity loop in the two-axis and four-gimbal airborne gyro-stabilized platform, the ADRC technology is applied in GSP control system. In the process of ADRC design, considering the actual system characteristics and the principle of easy to implement in engineering, the reference signal is directly introduced into the velocity loop system, that is to say, the arrangement for the transition process is a unit reference model. The PI feedback control which is commonly used has been chosen as our feedback control law in ADRC. In the design of Extended State Observer (ESO), we choose the linear function instead of non-linear function to simplify the design process which provides an efficient way to apply the ADRC on GSP control system.(2) In the design process of ESO, genetic algorithm(GA) is applied for parameter optimization. Since there is no perfect theoretical guidance in the design of ADRC parameters, empirical debug on control system is commonly used in engineering. In this dissertation, genetic algorithm toolbox is used for parameter identification after the stability conditions obtained in ESO:Using the genetic algorithm toolbox integrated in MATLAB to optimize the different fitness functions, four different sets of parameter values is obtained, which is quite efficient and quick and avoids the disadvantages of debugging based on experience.(3) The simulation platform is established and experimental analyses are made. Since the platform usually works at a low-speed, the non-linear friction torque is the main system disturbance. Based on the system model established by our research group in preliminary work, we built the simulation platform of the ADRC control system in Simulink. In simulation experiments, the estimation performance of ESO is evaluated whose parameters are obtained by the genetic algorithm toolbox. After comparison, the final parameter values are determined. After that, an experiment of estimated disturbance compensation in an open-loop system is done. The estimated disturbance by ESO is introduced into the input of the control system to offset the effects of non-linear friction torque. At last, experiments are done to analyze the performance of ADRC system and compare with the PI control system used in engineering by simulation.