Research On Multi-UAVs Cooperative Path Planning Based On Hierarchy Decomposition Strategy

With the rapid development of aerospace and information technology, coordinated operation ofmultiple UAVs will be the main conflict style in future war scenes. This paper is supported by theAviation Science Foundation (2009ZC52041). This paper is just to be dedicated to the research ofMulti-UAVs cooperative path planning.First, this paper constructed the UAV path planning model and useed some optimizationalgorithms to solve single path planning problem. In this paper, the UAV dynamic constraints andradar threat model were established, and respectively using the grid method and the Voronoi diagrammethod devided the environmental space. Then it expounded the traditional algorithm based onDijkstra algorithm and Ant Aolony Optimization to solve this problem. Through the in-depthsimulation analysis, it found the Voronoi diagram environment modeling method combined with AntColony Optimization algorithm can quickly obtained the solution. In order to further improve theconvergence speed and the results, an improved ant algorithm with direction guiding policy wasgenerated. Additionally，in order to meet the maneuverability performance constraints, three B-splinecurve is proposed to smooth the path.To solve the cooperative path planning problem for Multi-UAVs and inspired by McLain’sresearch, a decomposition strategy was proposed. This approach decentralized the large-scale,dynamic and high coupled optimization into three agents: path planning, cooperative manager andtrajectory smoother. Multi-UAVs coordinated combat model is presented, which took the space,precedence and time window constraints into account. In this paper, we discussed two attackstrategies, cooperative attacks and cooperative multi-tasks separately, and used the standard geneticalgorithm program to discuss the Task assignment/Path planning problem for Multi-UAVs incooperative manager. Because of Genetic Algorithm’s low efficiency, the parallel ant colonyalgorithm based on parallel computing idea was proposed to solve the problem.Finally, this paper discussed how to use the VR-Forces for the secondary development ofcomputer generated forces based on the visual simulation platform of the MAK series software. Thevisual simulation system for cooperative attacks and cooperative multi-tasks was developed on thisplatform.