Research On Fast Optimization Of Continuous Low Thrust Trajectory And Mission Decision Methods For Multiple Celestial Bodies Exploration

With the discovery of more and more celestial bodies,using limited propulsion resources and time to detect as many celestial bodies as possible has become a hotspot issue in the aerospace field.In order to reduce the launch cost and quickly explore more celestial bodies,using multiple spacecraft to detect a large number of celestial bodies with one rocket is also the development direction of deep space exploration in the fu ture.However,when considering the mission of continuous exploration of multiple celestial bodies,it is an urgent problem to choose which celestial bodies to detect and the detection sequence.At the same time,it is also very important to find a suitable transfer traject ory to save more fuel.However,with the increase of the number of detection targets in the detection mission,the complexity of the combination problem also increases.Therefore,for multi-spacecraft and multi-target detection missions,this paper mainly studies the fast optimization algorithm of continuous low-thrust spacecraft transfer trajectory,the fast selection algorithm of multi-target detection sequence,and the multi-spacecraft mission decision algorithm.The specific contents are as follows:For continuous low-thrust spacecraft such as ion electric propulsion spacecraft and electric sail spacecraft,the coordinate system and motion equations are given,the corresponding orbital dynamics model is established,and the thrust constraint conditions in the process of de ep space exploration are given.At the same time,for the gravity-assist transfer,the gravity-assist transfer model of spacecraft are given,which provides the basis for the subsequent analysis of trajectory optimization of continuous low-thrust spacecraft.The Bezier shape-based method is proposed for the rapid optimization of the transfer trajectory of continuous low-thrust spacecraft,and compared with finite Fourier series(FFS)method.The estimation process of state variables by the Bezier shape-based method is derived,the corresponding initialization methods of unknown coefficients are given,and the influence of unknown flight time parameters on trajectory optimization and unknown coefficients in the optimization process is analyzed.By comparing the Bezier shape-based method,finite Fourier series method and Gauss pseudospectral method(GPM),the superiority of the Bezier shape-based method in computational speed is confirmed.Aiming at the problem of multi-target continuous detection of single spacecraft,a fast selection algorithm of detection sequence based on Monte Carlo tree search(MCTS)algorithm is proposed.For different types of detection missions and propulsion systems,the corresponding mathematical models and constraints for continuous exploration of multiple celestial bodies are given.The selection process of optimal detection sequence based on Monte Carlo tree search algorithm is proposed,and the search process of synchronous calculation of detection sequence selection and trajectory optimization is introduced.The shape-based trajectory optimization algorithm is applied to the random simulation process of Monte Carlo tree search,and the fast calculation of the Bezier method is used to realize efficient searc h,and the construction process of search tree has practical physical significance;The selectio n of the detection sequence and the trajectory optimization are carried out synchronously.While directly outputting the current optimal detection sequence and the corresponding performance indicators,the accurate transfer trajectory is generated.Two improved Monte Carlo tree search algorithms,branch collection and bi-directional extended Monte Carlo tree search,are proposed to solve the mission decision problem in multi spacecraft and multi celestial body exploration.The mathematical model of mission decision-making in multi spacecraft and multi celestial body continuous exploration is established.The basic search principles and corresponding mission decision-making methods of two improved Monte Carlo tree search algorithms are given.Compared with the traditional Monte Carlo tree search algorithm,the two improved algorithms significantly improve the search efficiency,reduce t he difficulty of selecting the balance parameters and iteration times,and eliminate the possible repeated calculation in the traditional Monte Carlo tree sear ch algorithm.Compared with the two improved algorithms,the improved Monte Carlo tree search —bidirectional expansion algorithm has better se arch performance.