GA/DP Hybrid Solution for UAV Multi-Target Path Planning

摘要

Genetic Algorithm (GA) has made successful application in unmanned aerial vehicle (UAV) path planning. However, for multi-target problem, additional function should be included to enhance UAV performance. This paper introduces the Dynamic Programming (DP) for solving multiple target path planning into UAV system. The proposed hybrid solution integrates GA with DP for UAV to guarantee the safety requirement and feasibility for path planning. The Virtual Flight Map (VFM) is built for path planning algorithm. Between two targets, GA finds the optimal path with suitable fitness function; while with the distance matrices between all pairs of multiple targets, DP expands the two target path planning into multiple targets. The DP decides the visiting sequence of multiple targets and visits all targets just in once tour and finally returns to the starting point in the shortest route. The proposed algorithm is formulated and integrated for complete tests.