Path Planning For Unmanned Aerial Vehicle Based On Hybrid Algorithm

With the rapid development of artificial intelligence and UAV field.Many technique have been widely used in various fields of UAV,such as intelligent technology,navigation and positioning technology,motion control technology,multi-sensor fusion technology,multi-target collaboration technology and so on.In order to meet the requirements of UAV navigation task,path planning has become one of the important components of UAV.According to the circumstances that UAV perceive to the environment,it can be divided into local path planning and global path planning.This article is based on the rapidly-expanding random tree and dynamic window algorithms.Aiming at the shortcomings that the rapidly-expanding random tree algorithm has real-time performance and dynamic window algorithm is easy to fall into local minimums,the paper proposes a hybrid path planning algorithm which solve the path planning problem of the unmanned aerial vehicle in an unknown obstacle map.This article respectively introduces the rapidly-expanding random tree with restricted sampling area and the improved heuristic dynamic window algorithm.The core idea of the global path planning algorithm based on the rapidly expanding random tree is to take the starting point of path planning as the root node of the search tree,expand new nodes into the search tree through the constraints of ellipse bias sampling,and repeat the above steps until the target point is found.The basic idea of local path planning based on dynamic window algorithm is to sample the velocity space of the unmanned aerial vehicle,then predict the trajectory of the sampled velocity through the kinematics model of the unmanned aerial vehicle,and finally evaluate the estimated trajectory by using the evaluation function.(1)This paper proposes an improved heuristic dynamic window method to solve the problem of UAV with many turns and uneven paths.By limiting acceleration,dynamic constraints and adding curvature heuristic function,the turn times of UAV are reduced and the effectiveness of the algorithm is verified.Aiming at the low search efficiency and slow convergence speed of the fast-expanding random tree,an algorithm combining the artificial potential field method with the fast-expanding random tree is proposed.It makes the expansion node purposely expand to the target point.which improves the search efficiency and convergence speed of the fast-expanding random tree algorithm.(2)The purpose of this paper is a hybrid path planning algorithm for unmanned aerial vehicle in the case of the known global map information,through rapid expanded random tree algorithm for global path planning,it get information unmanned aerial vehicle Buddhism punctuation,reach each subtitle punctuation,respectively using dynamic window method to complete the path planning.The effectiveness of the hybrid path planning algorithm was verified by MATLAB simulation.Finally,the unmanned aerial vehicle completed the path planning in the unknown environment under the condition of short path length and smooth path,and successfully reached the target point.(3)By realizing the collaborative operation between MATLAB and Ubuntu/ROS computers,this paper takes the UAV as the simulation object under Gazebo physical simulation platform to carry out the simulation experiment of path planning.The hybrid algorithm of APF-I RRT and dynamic window method is used to make the path planning,and speed and position instructions are issued to the UAV to complete the entire path planning.The effectiveness of the algorithm is verified through the pipeline inspection task,and it is verified that the unmanned aerial vehicle under Gazebo simulation platform can realize autonomous obstacle avoidance,path planning,autonomous positioning and mapping under the dynamic uncertain environment.