Research On Reentry Trajectory Optimization And Guidance Method For Lifting Vehicle

Lift-type vehicle is a class of aircrafts with large lift-to-drag ratio,strong maneuverability,internal volume utilization,and good aerodynamic characteristics at high angles of attack.With the development of aerospace technology,the lift-type aircraft will be used in the future.It will play an important strategic role in politics and economy.The reentry segment is the process of re-entering the celestial atmosphere from the outer space by the lift-type vehicle.The research focus in trajectory planning and guidance.This paper takes this as the research object,and targets the bottleneck problem faced by the trajectory optimization and guidance of the reentry segment.Conduct research from the following aspects:Under certain assumptions,based on the kinematics and dynamics equations of the reentry segment of the aircraft,the three-degree-of-freedom centroid motion model and the atmospheric environment model of the reentry vehicle are established.Based on this,the subsequent trajectory optimization and guidance research are carried out.Aiming at the bottleneck of the traditional trajectory optimization algorithm in terms of fastness and convergence,a convex optimization method for online trajectory optimization is designed.The original trajectory optimization problem is a nonlinear optimal control problem,which is convex to SOCP.The sub-problem form is separately convex from the aspects of motion equation,performance index,process constraint and problem discretization,and the convex problem after processing is solved by the originaldual interior point method.In order to verify the performance of the trajectory optimization algorithm in terms of fastness and convergence accuracy,this paper is based on the reentry vehicle model,and uses this method to solve the trajectory optimization problem under the shortest time and the minimum voyage respectively.The simulation results show the effectiveness and feasibility.In the process of re-entry trajectory optimization,most of the algorithms have strong dependence on modeling,which leads to the problem of poor adaptability of the model under the existence of deviation or disturbance.A trajectory optimization intelligent decision combining reinforcement learning and neural network is proposed.algorithm.The traditional trajectory optimization algorithm mostly solves the control quantity by numerical method.In the method proposed in this paper,the action value is mapped by the neural network as the state quantity for the state of the aircraft,and the reward is brought by the state transition.Learning training mechanism,training neural network;based on the reentry vehicle model,this method is used to solve the feasible orbits satisfying process constraints and terminal constraints,and verify the feasibility of the method.With computer computing performance and intelligent decision system increasing,this method will have a better application prospect in the complex continuous control problems such as trajectory optimization.Aiming at the improvement of real-time and convergence requirements of the guidance system,a reentry guidance method based on quadratic constrained quadratic programming(QCQP)is designed.The guidance method uses the tracking error of the reference trajectory and the terminal falling point error as the optimization performance index and convex constraint,and converts the nonlinear optimal control problem corresponding to the original form of the guidance problem into the QCQP problem,using the interior point method.The problem is solved quickly by the polynomial time complexity,and the optimal guidance law is obtained.In this paper,the validity of the verification algorithm is verified with the initial state deviation,the aerodynamic parameter deviation and the atmospheric density deviation interference.Finally,the paper summarizes the whole thesis,points out the innovation of this paper,and the direction and improvement of research in the future.