The problem of identification parameter data saturation in repetitive control and its solution

Abstract

In the repetitive control of tracking periodic signals based on the principle of internal model, the control effect has a great relationship with the parameters of the controlled system. If the system is affected by noise and causes the internal parameters to change, failure to obtain the repeated control of the internal parameters in time will cause the system to lose stability. Therefore, how to quickly identify the parameters of the controlled system is particularly important in the field of repetitive control. In the actual process, the traditional least square method is often used to identify the parameters of the controlled system. However, the convergence of the algorithm to parameter identification is very slow. Once the controlled system parameters are changed, the parameter information provided by the new data cannot be updated in time, and the convergence of the identification results is very slow. In order to overcome the data saturation phenomenon of the least squares algorithm, this paper uses three methods of forgetting factor algorithm, variable gain matrix algorithm, and introducing additional matrix R algorithm to improve the traditional least squares identification algorithm, and verified these three through MATLAB simulation. Effectiveness of the method. Compared with traditional methods, the improved three identification methods can speed up the convergence of parameter identification and improve the accuracy of parameter identification.