Parameter Tuning Methods Of Linear Active Disturbance Rejection Control For Radar Position Servo System

Airborne radar plays a very important role in achieving target tracking,air alert,air reconnaissance and so on,so it occupies an important position in aviation,navigation,military and other fields.The control performance of radar servo system directly affects the real-time target detection performance,accuracy and stability of airborne radar,so the research on radar servo control system is of great significance.There are two main difficulties in the design of the radar servo control system: on the one hand,the airborne radar has a harsh working environment and is highly susceptible to environmental influences such as wind direction,air flow,temperature and so on.At the same time,the vibration of the body and the attitude of the carrier will also affect its stability and accuracy;on the other hand,there are inevitably nonlinear factors such as component friction and backlash dead zone in the airborne radar servo system due to gear rotation,so its mathematical model is generally difficult to establish accurately.If reasonable control strategies are not applied to such a nonlinear time-varying system,the airborne radar will not work normally.Therefore,this subject will study the radar servo active disturbance rejection control system and optimize its parameters,and finally achieve the goal that the airborne radar can maintain good performance with lower control energy.The main research content of this paper is as follows:Firstly,the radar position servo system has been taken as an example to derive the mathematical model,taking the components friction and the environmental disturbance into account.At the same time,the traditional active disturbance rejection control(ADRC)and linear active disturbance rejection control(LADRC)are both designed for the radar position servo system.Due to the convenience of LADRC parameters tuning and stronger practicability,this paper mainly focuses on the linear active disturbance rejection tracking control of the radar position servo system.The MATLAB simulation results show that,compared with the traditional PID,the LADRC can effectively suppress internal nonlinear disturbance and external environmental interference,and can also reduce the dependence of algorithm on the system model.However,the control accuracy of the control system needs to be improved in order to realize the ultra-high-precision operation of the radar position servo system,and the control system has the problem that it is difficult to adjust and optimize the parameters online.Secondly,an adaptive law based linear active disturbance rejection control(ALADRC)is designed aiming at the problem that the control system has low tracking accuracy and the control parameters are difficult to be adjusted online.The adaptive law can make the control parameters automatically adjust themselves to achieve a certain control effect,and to adapt to the dynamic characteristics of the radar position servo system.The MATLAB simulation results show that this composite control method can compensate the tracking error caused by the deviation of the control parameters,not only achieving higher control precision,but also enhancing the anti-interference ability of the controller to achieve the optimal control effect compared with traditional LADRC.Thirdly,the BP neural network based linear active disturbance rejection control(BPNN-LADRC)is further proposed aiming at the problem that the parameters tuning of ALADRC is slow and the control accuracy is slightly improved.With the ability of BPNN to map and approximate functions,the optimal solution of control parameters can be obtained through self-learning of the fastest gradient descent algorithm.The MATLAB simulation results show that BPNN-LADRC has higher control accuracy,smaller tracking error,and system control performance has been further improved compared with ALADRC.Fourthly,the diagonal recurrent neural network based linear active disturbance rejection control(DRNN-LADRC)is further proposed aiming at the problem that the BP neural network cannot well control the dynamic time-varying system due to its static characteristics.DRNN is a dynamic network with the feedback structures and varying degree of delay in its hidden layers,so it has certain memory and learning capabilities,and is more suitable for the identification and control of radar position servo dynamic system.The MATLAB simulation results show that DRNN-LADRC has better control performance compared with BPNN-LADRC.Fifthly,three nonlinear functions,namely “interval function”,“sawtooth function”and “gradual step function”,are constructed and added to the radar position servo system,and the Gaussian white noise interference is added as well,in order to effectively verify the robustness of the above mentioned composite controller and these applicability in practical applications.The MATLAB simulation results show that the robustness of the three proposed composite controllers is satisfactory.