Research And Implementation Of Quadrotor Motion Planning Technology For Unknown Complex Environment

Motion planning is a core module of the autonomous navigation system of quadrotor in unknown and complex environments.Motion planning is responsible for generating a state transition trajectory from the starting position to the target position for the UAV.It can be widely used in military or civilian scenarios such as battlefield reconnaissance,fire strike,aerial photography,search and rescue in disaster,and autonomous inspections.and provide effective support for these application tasks.At present,the main challenge of motion planning in unknown complex environments is how to improve the safety of the UAV’s flight trajectory under the constraints of limited sensing range without sacrificing navigation efficiency.To this end,the paper focuses on the problem of quadrotor motion planning in unknown and complex environments,and proposed a topological waypoint search method based on ESDF gradients and a conservative adaptive trajectory re-planning mechanism,and combines them to build an autonomous navigation system of quadrotor for unknown complex environment.On this basis,the research results of this paper have been verified by experiments.The main contributions of this paper include the following three aspects:(1)Aiming at the problems of global known space search,path security and local optimal trap in classical path searching algorithm,an ESDF gradient-based topological waypoint search method is proposed.The core of this method is to extract some feasible topological path points from both sides of the obstacle by using the geometric information of the obstacle surface and the ESDF gradient at the collision point,thereby reducing the search space.At the same time,in the waypoint search process,the distance from the path point to the obstacle is taken as an important parameter,which effectively improves the safety of the geometric path and can better deal with the large obstacle scene.Finally,according to the real-time motion trend of quadrotor combined with dynamic factors,the path point with the lowest transfer cost is selected as the next waypoint.(2)Aiming at the problem that the existing optimistic motion planning algorithms require frequent trajectory collision detection and re-planning to ensure the safety of quadrotor,a conservative adaptive trajectory re-planning mechanism is proposed,which can ensure the trajectory safety and reduce the frequency of re-planning.Firstly,the unknown environment beyond the quadrotor’s field of view is conservatively regarded as impassable,and only the trajectory segment located in the safe area within the quadrotor’s field of view is optimized to ensure the safety of the trajectory.Furthermore,a trajectory re-planning trigger mechanism based on unsafe events is proposed,which plans to generate a new safe trajectory only when the UAV detects that the current trajectory will collide or is close to the obstacle.Finally,the local trajectory is parameterized by B-spline curve to improve the efficiency of trajectory optimization process.(3)Based on the above research results,a quadrotor autonomous navigation prototype system for unknown complex environment is designed and implemented,the experimental environment was built,and the maps containing different obstacles are constructed in the ROS platform.The topological waypoint search method and trajectory re-planning mechanism proposed in this paper are experimentally verified.The results show that compared with the existing methods,the method and mechanism proposed significantly improve the trajectory security and planning efficiency,and have a certain degree of map universality.