Research On Large Time Delay Control Method Of Hypersonic Vehicle Based On Neural Network

In order to solve the problem of large delay and input constraints in a hypersonic vehicle input,this paper makes an exploratory research on the design and improvement of the control algorithm.Large delay and input constraints can make the control quality decrease,resulting in long regulation time,the big overshoot,and even the instability of the system.These defects make the control algorithm difficult to satisfy the control requirements of a hypersonic flight vehicle,and even can make the system out of control under the complex environment.Therefore,it is significant to study the control method of hypersonic vehicle with large time-delay and input constraints in engineering.In this paper,the dynamic equation of longitudinal channel of a hypersonic aircraft is established by analyzing the simple force of the aircraft in the longitudinal channel.This model is a nonlinear and strong coupling variable system.It is difficult to establish the accurate mathematical model in the practical application of control,so it is almost impossible to analyze and study the algorithm through the mathematical model,which presents a new challenge to the research of control algorithm.Based on the established aircraft model,this paper firstly selects the traditional PID control algorithm to study the tracking control process of the Maher number.The simulation results show that the function of the conventional PID control method can not reach the steady state.At the same time,the control system is a slow system.Under the function of traditional PID control method,the regulation time is too long and the overshoot is too large to meet the actual control demand of hypersonic vehicle with the input constraint.On the basis of a large number of experimental data,through the analysis and theoretical proof,the input constraints are taken into account in the design of control algorithms,and a new incremental PID control method which is based on input constraints is proposed.The simulation results show that the new incremental PID control method can solve the control problem of large overshoot and long adjustment time in the original control system.In order to solve the instability of the system caused by the large time delay,this paper utilizes the capability of nonlinear approximation of BP neural network to tune the conventional PID parameters online to achieve the purpose of nonlinear combination of the three parameters.The simulation results show that the PID control method based on BP neural network can effectively avoid the equal amplitude oscillation,and the control system can reach the stable state.Although the BP neural network PID control method can effectively avoid the amplitude oscillation,but the control method can not guarantee the accuracy of the control system,the final Maher number will appear relatively large steady-state error.For the precision requirement of the control system,a new PID control method based on neural network is proposed in this paper.The simulation results show that the algorithm can effectively eliminate the static error of the system.However,the simulation results show that the convergence speed of the control system is very slow and can not meet the needs of rapid response.In order to meet the demand of fast response of control system,this paper improves the original neural network learning algorithm.The constant learning algorithm is transformed into an adaptive learning algorithm.The simulation proves that the modified control method can effectively improve the response speed of the aircraft.