Research On Indoor 3D Path Planning Method For Autonomous Navigation UAV

With the rise of indoor navigation and positioning technology,the application of UAV in indoor environments has been unprecedentedly developed.Compared with the general UAV,quadrotor UAV has the advantages of small volume,flexible maneuverability,simple control and easy take-off.It can replace human beings to perform indoor target search,harmful gas information inspection and other tasks,and has broad application prospects in anti-terrorism investigation,post disaster search and rescue and other fields.Due to the complex indoor space structure,limited height and many obstacles,UAV has great challenges in indoor autonomous navigation.As one of the key technologies of UAV autonomous navigation,path planning has become a research hotspot in the field of UAV This paper focuses on the indoor autonomous navigation of quadrotor UAV,and studies the three-dimensional path planning method of quadrotor UAV.The main research work of this paper is as follows:(1)Aiming at the trajectory planning of the indoor quadrotor UAV;firstly,the existing environmental modeling methods are analyzed,and the grid modeling method is used to realize the discrete processing of the indoor continuous planning space;then combined with the quadrotor UAV itself flight characteristics,analyze the physical limitations of its flight in an indoor environment;researched common algorithms in path planning problems,and selected reinforcement learning algorithms and A*algorithms as the research objects of the trajectory planning method in this article by analyzing the advantages and disadvantages of different algorithms;(2)In view of the complex indoor space environment and the long convergence time of the Q-learning algorithm,a space optimization strategy for simplifying algorithm search nodes,and a Q value initialization strategy for determining the target direction relative to the starting position are proposed.The experimental results show that:compared with the traditional Q-learning algorithm,the improved Q-learning algorithm has a significant reduction in convergence time and the number of search nodes,and the planned path length is also shortened,which is verified in different environments Improved algorithm has good generalization ability;(3)In view of the problem that the track planned by the traditional A*algorithm is close to the obstacle,the safety distance function is added to the heuristic function,and the planned track is smoothed;in order to meet the actual needs of the quadrotor UAV indoor navigation,re-planning and processing the sudden obstacles that appear on the flight path of the UAV.The experimental results show that:compared with the traditional A*algorithm,the track planned by the improved A*algorithm can be far away from obstacles to a certain extent,ensuring the flight safety of the UAV,and the smoothed track is more in line with the actual flight of the UAV;the re-planning of sudden obstacles appearing on the global trajectory can better meet the real-time requirements of UAV.