Research And Simulation Of Multi-Dimensional Taylor Network Optimal Control For The Flight Of The Plane

Aircraft is one of the most important inventions of the twentieth Century, in which the military aircraft is an indispensable equipment of the modern battlefield. The fighter is the main variety of military aircraft, so all countries have put the development of fighters in a crucial position. In order to achieve flight precise control and manipulation, high quality flight control system should be met, which is an indispensably significantt part of modern aircraft design technology, but also an important guarantee for flight safety and the complex flight tasks. Kinematics model of aircraft has the characteristics of strong coupling, nonlinear and the complexity of the mechanism. The newly proposed Multi-dimensional Taylor Network optimal control algorithm is adopted to design the controller in the thesis, also the sliding mode variable structure control and PID control is used for comparing, the sliding mode variable structure control is mainly introduced, also control methods are simulated and compared in Matlab.According to the physical model of the aircraft, nonlinear, fully coupled six degree of freedom motion equations of aircraft is established under the ideal condition, taking into account the actual flight, the limiting the steering gear and the gyroscope structure are joined in simulation.Firstly, the sliding mode variable structure control is selected as the first control method. In this thesis, reaching law sliding mode is used to design the controller in saparated channels. The longitudinal channel contains the pitch angle and height control, lateral channel includes rolling control and yaw control. The loop pitch angle and roll angle based on reaching law sliding mode design of sliding mode controller to ensure the robustness of the system, and combining with the outer loop PID algorithm of height tracking and heading control to complete the design of the flight control system. Lateral control includes horizontal turning and coordinated turning.Then MTN(Multi-dimensional Taylor Network) optimal control which is newly proposed is selected as the second control method. This thesis introduces the theory and application of MTN optimal control, then introduces the self-learning algorithm for obtaining optimal control signals using the minimum principle. Controller is designed in separated channel. MTN optimal control designs the inner loop controller, and the parameters are optimized by using the conjugate gradient method, and the outer loop is controlled by PID algorithm. MTN parameters are optimized by conjugate gradient method, and the simulation experiment is done in separated channels.The controllers designed in separated channels are combined. The whole channel simulation is carried out, and the control effects of two kinds of control algorithms are compared. The wind disturbance model is introduced, and the model is added to the mathematical model of the aircraft, and the control effects of the two control methods are compared. Flight control system simulation platform is designed based on GUI with MATLAB. Under the same conditions, the simulation results show that the MTN optimal control has a good control effect. Compared to sliding mode variable structure control and PID control, using MTN optimal control as the controller of aircraft flight has strong maneuverability and anti-interference.Finally, what have done is concluded, at the same time the outlook for further follow-up work is offered.