Research On Lunar Rover Path Planning Algorithm Based On Slip Prediction In Complex Terrain Environment

Lunar rover refers to a special vehicle that can move autonomously on the surface of the moon,detect scientific targets,and conduct scientific patrols.It can also be called a lunar patrol.The path planning algorithm is a key technology in the autonomous navigation system of the lunar rover.Whether it can plan a safe and collision-free path to the target point is an important guarantee for the lunar rover to conduct scientific investigations and experiments.With the development of lunar exploration projects,in order to detect higher-value scientific targets,the future planning environment may have more complex and dynamic environments,and the factors that need to be considered will gradually increase.Therefore,in view of the failure of path planning caused by the sliding of the traditional planning algorithm in soft terrain,this paper improves the global planning and local planning respectively.Among them,the cost function of slip prediction is added to the global planning algorithm D*Lite,so as to avoid high slip areas on a large scale.In local path planning,because the environment is completely unknown or dynamic,the traditional method based on pre-set rules will not be able to quickly and accurately predict the cost of the path due to the lack of generalization ability.This paper applies the reinforcement learning method in machine learning to solve the path planning problem,and proposes an end-to-end local path planning algorithm.Two networks are designed for predicting cost and output action respectively,so that the influence of slip is added to the consideration of the algorithm.At the same time,by designing a reasonable environment and reward function,the algorithm’s dynamic obstacle avoidance ability in a complex terrain environment is improved,and the effectiveness and generalization ability of the algorithm are verified in simulation experiments.The experimental results show that the lunar rover path planning algorithm based on slip prediction in the complex environment proposed in this paper can effectively avoid the high slip area and produce a trajectory with lower deviation from the expected path,thereby effectively improving the safety of path planning.Improve the learning ability of the algorithm and reduce the cost of planning the path.