Research On Path Planning And Task Allocation Of UAV In 3D Complex Environments

With the development of artificial intelligence technology,the performance and autonomy of unmanned aerial vehicle(UAV)are also improving.At present,UAV has been widely used in the military field,and multi-UAV cooperative operation has become an important form of UAV operation in the future.Path planning and task allocation are two core issues of multiUAV cooperative mission planning,which is of great significance to improve the combat effectiveness of UAV group.Based on the research background of UAV performing military tasks,mathematical models are established for off-line path planning of single UAV,on-line path planning of single UAV and cooperative mission planning of multiple UAVs.A variety of methods are proposed,and then the performance of these methods is verified and compared through example tests.The test results show that these methods can provide feasible and effective solutions for UAV path planning and task allocation.In summary,the main research work of this paper is as follows:(1)For the off-line path planning problem of single UAV,a complex 3D environment model is established.Taking minimizing path length and minimizing radar threat as optimization objectives,two genetic algorithms(GA)with different population initialization methods and genetic operations are proposed.Experimental results show that both algorithms generate good feasible solutions.When the task execution environment tends to be complex,the genetic algorithm based on zone initialization can obtain better solution effect.(2)For the on-line path planning of single UAV,a dynamic obstacle model is designed.Based on the pre-planned path obtained offline,rapid exploring random trees(RRT)algorithm and artificial potential field(APF)algorithm are proposed to replan the path online for UAV.The results show that the two proposed algorithms can meet the real-time obstacle avoidance requirements of UAV.In addition,there is no significant difference in the solution quality between the two algorithms,but the computational efficiency of APF algorithm is better than RRT algorithm.(3)For the multi-UAVs cooperative mission planning problem,a combat scenario in which multiple UAVs perform multiple tasks and then assemble is defined.Based on the objective of maximizing the efficiency of all tasks,a mathematical model describing the problem is established.Then two task allocation methods based on particle swarm optimization(PSO)and consensus-based bundle algorithm(CBBA)are proposed,and a collaborative strategy is designed.The results show that the combination of the two algorithms and collaborative strategy can obtain a feasible mission planning plan with equivalent effect.