The Research Of UAV Intelligent Path Planning Strategy In 3D Dynamic Environments

With their high mobility,low cost and flexible configuration,unmanned aerial vehicles(UAVs)are projected to have an enormous impact on revolutionizing applications ranging from air transportation,mobile communication,detection,and monitoring.Path planning is a worth-wondering issue for the full deployment of UAVs,as well as providing reliable services to fulfill diverse missions.Currently,there are already a large number of researches and results considering the UAV path planning problems in various scenarios.However,for the simplified constraints and environmental modeling,they still have poor performance in the real-time calculation and dynamic obstacle avoidance.Considering the trade-off between the planning speed and accuracy,the effective solution is the combination of the off-line and on-line trajectory design for obstacles and potential collisions avoidance in dynamic environments.Correspondingly,the research mainly deals with problems of the globally optimal path for static obstacles and local trajectory design for moving obstacles.The major challenges consist of two aspects:1)Due to the aerodynamic and environmental constraints of UAVs,the computational complexity is too large to ensure the accuracy of path planning in the large-scale static environments;2)The complex process of environment modeling,collision prediction,and collision avoidance are included in the trajectory design for moving obstacles.It requires a long computation time,which can hardly ensure the flight safety of UAVs.To deal with the challenges,this research proposes the combination of static global and dynamic local intelligent path planning methods to solve the problem of UAV path planning in 3D dynamic environments.The main work includes:1)The comparison and analysis are presented for the performance of existing path planning methods in system modeling and path searching;2)The VC-JPS method is proposed to reduce the complexity in large-scale static environments,the cell-varying strategy in hexagonal cells makes the values of dynamic parameters such as yaw rate and pitch rate validated to meet the constraints of UAVs,and the pruning mechanism,jumping rules are also improved to decrease the computation time;3)The VC-VQ intelligent trajectory design strategy based on hybrid dynamic trajectory planning methods is proposed for the reduction of computation time and improved the obstacle avoidance rate through varying velocity and reinforcement learning methods.