Research On Active Obstacle Avoidance Technology For UAVs In Dynamic Environments

Unmanned Aerial Vehicles(UAVs)have demonstrated their potential in various military and civilian scenarios.Real-world situations often involve dynamic and irregular obstacles,necessitating UAVs to avoid obstacles effectively to ensure safe flight.This thesis focuses on motion planning for obstacle avoidance in dynamic environments and analyzes several critical aspects,including path search,flight corridor generation,trajectory optimization,and emergency obstacle avoidance planning,to develop a motion planning system for obstacle avoidance.The thesis’ s main contributions are summarized as follows:1.A fast path search algorithm is proposed to address the problem of path search algorithms not fully utilizing known obstacle information,leading to long search time.The algorithm employs an adaptive strategy to determine the optimal inflation factor of the current map and utilizes this optimal inflation factor to perform an improved Hybrid A* algorithm.Three research innovations,namely heuristic functions,occlusion penalty terms,and adaptive strategies,are emphasized.Experiments reveal that the proposed algorithm can balance path length optimality and motion target observation,effectively plan optimal paths for UAVs in complex scenarios,and reduce path search time.2.A flight corridor generation algorithm based on the endpoint arbitrary straight-line grid calculation method is proposed to tackle the problem of inaccurate straight-line grid calculation and long generation time of existing flight corridor algorithms.The algorithm first calculates the initial flight corridor,determines the corridor boundary using the line with arbitrary endpoints grid calculation method,searches for obstacle grids within the boundary,and updates the flight corridor.The proposed algorithm significantly reduces the flight corridor generation time.3.A dynamic obstacle avoidance modeling and trajectory optimization algorithm is proposed to address the problem of long and difficult trajectory solutions in existing obstacle avoidance modeling methods due to the introduction of non-convex constraints and more auxiliary variables.The algorithm classifies dynamic obstacle trajectories into three categories based on their positions in the flight corridor and models the optimal trajectory as a convex quadratic programming problem based on interval boundary constraints and instantaneous obstacle avoidance constraints.Simulation experimental results demonstrate that the algorithm performs well in all representative dynamic scenarios.4.An emergency obstacle avoidance planning algorithm is proposed to address the problem of jitter and local minima in current emergency obstacle avoidance algorithms.The algorithm models the obstacle wall equation and designs a smooth obstacle avoidance trajectory to achieve advanced avoidance of the obstacle.Experimental results confirm that the algorithm can make the UAV move away smoothly and quickly when it is close to the obstacle,thereby enhancing the UAV’s obstacle avoidance capability in dynamic scenarios.