Application Of Adaptive Dynamic Programming In Path-planning And Optimal Control Problem Of Space Engine

Approximate dynamic programming is an effective method to solve complex nonlinear optimal control problem with unknown model.ADP is helpful in solving the problem of" curse of dimensionality",which is difficult to overcome for dynamic programming.ADP calculates the optimal control realtime.In this paper,we focus on two classes of optimal control problems with complex nonlinear dynamic systems.The first problem is to design a optimal path-planning algorithm for mobile robots.The path planning of mobile robot is usually realized by two kinds of algorithms.Based on the prior static obstacle information,the mobile robot first obtains the global reference path by the global path planning algorithm.In the process of mobile robot motion,the final real-time path is obtained by local path planning algorithm based on the dynamic obstacle information obtained by sensors and global reference path.The adaptive dynamic programming does not depend on the specific model of the system.It has strong real-time performance and fast planning speed.Based on these advantages,an optimal control algorithm using ADP to select local target points is proposed,so as to improve the performance of local path planning algorithm.The modified algorithm can optimize the path length.Since the first algorithm may increase the time cost of path planning,this paper continues to propose a path planning algorithm based on ADP and artificial potential field method.Based on the known global obstacles,the simulation experiment of this paper investigates the dynamic obstacle avoidance and path-optimization ability of the mobile robot under the guidance of the proposed algorithms.The second problem is to solve the optimal control problem for space engines.It is difficult to get the accurate model for the space engine,which has a large complex nonlinear dynamic characteristic.The space engine researchers focus on how to get the model of space engine and how to control it.In order to get the perfect control for space engine,the algorithm should guarantee the stability,real-time performance and safety.In this paper,we research on improving the ADP algorithm to get the optimal control for space engines.The simulation experiment in this part is based on the model in simulink with Matlab 2020a provided by the Sun Ximing space engines studying team.