Research On Large-scale Secure And Complete Path Planning Method Based On Lunar Path Topology Network

With the development of the lunar exploration mission from simple patrol detection to more complex resource development and utilization and lunar base construction,the capability demand for the lunar rover has also changed from local small-scale detection to large-scale detection.The large-scale autonomous mobile detection capability is a key capability to realize the tasks of lunar polar region resource exploration,wide area sampling return of lunar surface and lunar base construction.The current research on path planning of lunar rover is mostly local path planning with small scope and accurate constraints,but there is still a lack of systematic and complete technical system for large-scale mobile path planning.In view of the above background,this paper creatively puts forward the construction idea of lunar road network,and studies three main problems: lunar road topology network design method,comprehensive path planning method based on lunar road topology network and patrol intelligent real-time path planning method with security constraints.The main research results include:The relevant theoretical models involved in the design of lunar road topology network and the application of large-scale mobility planning are modeled and analyzed.Firstly,according to the basic theory of complex network,the structure and function indexes and basic network model required for network design are given,and the algorithm completeness of planning based on path network is analyzed,which provides theoretical support for the network design and application of lunar path topology network.Then,the relevant models of lunar rover movement planning are established,including lunar terrain feature extraction model based on lunar digital elevation map,rover movement model and lunar feasible area analysis model,so as to provide model support for subsequent network construction and path planning.A lunar path topology network design and optimization method based on the research idea of ‘discover-model-analyse-optimize’ in complex network theory is proposed.Firstly,the design requirements and basis of lunar path network are analyzed.Then,the basic model design method of lunar road topology network is proposed,so that the generated network basic model meets the set design constraints and has the basic function of network topology.Then,according to the relevant indexes of complex network,the topological structure characteristics and functional characteristics are quantitatively evaluated.Finally,the optimization design method of lunar road network is proposed.The optimization design of the network model is realized through the design optimization model and optimization algorithm,so that the network can achieve better road topology and traffic capacity.A comprehensive path planning method for large-scale detection based on lunar road topology network is proposed.Firstly,a multi-cost comprehensive fast path search algorithm is proposed.Through the comprehensive analysis of lunar terrain,illumination and other environmental constraints,a multi cost heuristic function is designed,and a fast heuristic search strategy is designed,so that the algorithm can comprehensively consider the constraints of the complex lunar environment and realize the fast search of a wide range of paths.Then,a K-optimal path planning method based on lunar road topology network is proposed,realized more complete and efficient lunar large-scale mobile path planning.Finally,a fast re-planning method based on local topology strategy is proposed to achieve more robust and efficient path re-planning through high confidence passable map and local topology strategy.An intelligent real-time path planning algorithm with security constraints based on deep reinforcement learning is proposed.Firstly,in order to eliminate the difference between training environment and real environment,a simulation training environment integrating real lunar surface elevation data is built.Then,an end-to-end intelligent path planning algorithm with security constraints is proposed,which can realize that the inspector can directly make safe movement planning action only by relying on sensor data.Finally,the effectiveness of the model in this chapter is verified by simulation test,the comparison with the classical algorithm shows that the path generated by the algorithm proposed in this paper has higher security guarantee.To sum up,based on the construction assumption of lunar road topology network,this paper focuses on the problem of large-scale safe mobile path planning of lunar rover,in order to realize more safe,complete and efficient lunar large-scale mobile path planning through the design and construction of lunar road topology network,and provide valuable development and construction ideas for China’s future lunar exploration missions.