Design Of Adaptive Tracking Control System Based On Support Vector Machine Regression

As the control system is becoming increasingly complex,production practice and science and technology are having an increasingly higher requirement on the performance of the automatic control system,Adaptive Control Theory has come into being.In adaptive control,as the parameters of the controlled objects are frequent variables,system models can be identified online and timely adjusted,therefore,the system operation stability and reliability can be guaranteed.It has not only been successfully and widely applied to the field of classic control areas,but also in other intelligent robots,smart home and other emerging fields of science and technology in control research,which is also seen to have made a great progress.In this paper,adaptive tracking control is applied to network refrigeration system platform,and the adaptive control design of network refrigeration system is processed at the inception.Network refrigeration system platform is built,and the model of network refrigeration system is established through physical modeling method,and then the adaptive controller of the network refrigeration system is designed.Simulation results show that the tracking effect of the controller is not ideal when the system parameters remain constant.In order to improve the control accuracy,a SVR-based online adaptive tracking control method is presented in this paper.By using the SVR-based identification algorithm,a real-time tracking and processing are implemented to the three parameters of the system model.Firstly,through using the SVR-based algorithm to implement offline identification of the system model parameters,whose regression forecasts are realized through variables selection and data normalization;secondly,through using Support Vector Regression online identification model parameters,the adaptive tracking controller is adjusted based on parameter variations;finally,the simulation results demonstrate the SVR-based adaptive tracking control method has an ideal control effect.