Manned lunar landing track design method, device and equipment based on reinforcement learning

Abstract

The invention relates to a manned lunar landing orbit design method, device and equipment based on reinforcement learning. The method comprises the steps that task constraint conditions of an established full-task orbit optimization model are acquired; obtaining a decision action space of the intelligent agent; setting a Markov decision model and parameters of a Q learning algorithm in reinforcement learning according to the task constraint condition and the decision action space; setting a speed increment threshold value in a reward function of a Q learning algorithm according to the total income of the manned lunar landing full task; a reward function of the Q learning algorithm is the sum of an original reward function and a heuristic reward function; starting optimization calculation of a Markov decision model and a Q learning algorithm until an optimal flight scheme of the intelligent agent is obtained; the optimal flight scheme comprises target orbits for intelligent agent departure and arrival; and according to the optimal flight scheme, performing extrapolation calculation by using a real orbit dynamics model to obtain a full-task orbit scheme of the intelligent agent. And manned lunar landing full-task multi-stage track optimization and design are realized.

Description

technical field [0001] This application relates to the technical field of aerospace track design, in particular to a method, device and equipment for manned moon landing track design based on reinforcement learning. Background technique [0002] The moon is the only gateway for human beings to go to deep space. With the development of aerospace technology, the conditions for carrying out lunar exploration missions are becoming more and more mature. The manned lunar exploration mission is a complex system engineering. The whole mission consists of multiple flight stages, and there are strong nonlinear constraints in different stages and between stages. One of the core issues is the research on the design and optimization of manned lunar exploration orbits. The earth-moon space transfer orbit for manned landing on the moon is mainly divided into two categories: the earth-moon transfer orbit and the moon-earth transfer orbit. Existing track design techniques mainly include co...

AI Technical Summary

Problems solved by technology

However, in the process of realizing the present invention, the inventors found that the existing design and analysis methods for manned lunar exploration orbits cannot be used for spacecraft to perceive the environment and make decisions at different flight stages, and there is a problem that the manned orbit cannot be carried out. Technical Problems of Multi-stage Orbit Optimization and Design for the Full Moon Landing Mission

Images