Research On UAVs Path Planning Based On Passive Localization

The past decade has witnessed the rapid progress of the technology of the UAVs in military and civilian fields.In particular,because of its small size,strong maneuverability,good concealment,UAVs have been widely applied in electronic reconnaissance systems.With the development of UAVs,UAV’s engagement in the cooperative battle becomes a major development trend of future UAVs combat methods.The ability to detect and locate is crucial for reconnaissance UAVs.Therefore,this paper studies the UAVs path planning based on passive localization,which is of great significance for improving reconnaissance performance,accelerating task execution speed,and enhancing combat efficiency.The main research content of this paper is summarized below:To deal with the problem that the formation of UAVs influenced the passive localization accuracy in the electronic reconnaissance system,we introduced the Cramer Rao Lower Bound(CRLB)into UAVs path planning to optimize the passive localization accuracy along with distance cost,threat cost,kinematic constraints and distance constraints.Passive localization accuracy is thereby ensured by maintaining a good formation during the UAVs movement.To deal with the problem that the traditional intelligent optimization algorithms are used in complex environments for path planning,they usually have low optimization ability and tend to fall into the local optimal problem.To overcome the above limitations,a multi-objective grasshopper optimization algorithm(MOGOA)based on adaptive search is studied to plan the three-dimensional path of UAVs.Searching for the best location in search space by simulating the process of locust predation,and the Pareto frontier is used as the elite population to increase the selection range of search center and improve the diversity of population,and then the change of search range is adjusted adaptively by the success rate.Finally,the results show that the proposed method is able to plan a reliable path and ensure the accuracy of passive localization during UAVs flight;moreover,compared with the original GOA and MOGOA,it has a higher convergence speed and better global search performance.The multi-agent deep deterministic policy gradient(MADDPG)algorithm is introduced into the research on UAVs path planning based on passive localization.To deal with the problem that excessive training leading to difficulties in ensuring real-time performance when path planning is performed during training.A three-dimensional UAVs path planning method is proposed,which combines IMOGOA with MADDPG.First,MOGOA is used to search the optimal position of the UAVs,thus obtaining the initial path and accelerating the training during the initial training;in the subsequent training process,UAVs can be trained based on the previous training network.Finally,the results show that the method proposed in this section can effectively improve the algorithm’s training speed and obtain a reliable path.