Research On Path Planning And Autonomous Obstacle-avoidance Algorithm For Unmanned Vehicle

Autonomous navigation technology is the core technology of unmanned ground vehicles,which is a hot-spot issue of research in the field of artificial intelligence.Generally speaking,the autonomous navigation of unmanned ground vehicles consists of four parts: perception,localization,path planning and control.The path planning problem,as an integral part of unmanned ground vehicles research,has greater research and application value.Although many scholars have proposed various algorithms to analyze and solve this problem,there are not many effective methods,which is the necessity of this paper to continue to study the path planning problem.The contents and research results of this paper are as follows:(1)A global path planning algorithm based on the improved A* algorithm is implemented.Compared with the traditional A* algorithm,the algorithm adds directional cost to the cost function,reduces unnecessary grid search and the number of turns,and proposes an approximate Euclidean distance calculation method to make it faster than the exact value.Meanwhile,the Weight A* algorithm is used to improve the convergence rate of the algorithm when it is close to the target.Then,the planned path is smoothed by using the cubic spline interpolation,which can make the finally obtained path more reasonable.According to the application needs of ground unmanned vehicles in autonomous cruising,an autonomous cruising scheme based on virtual channel is proposed by drawing on the idea of Safe Flight Corridor(SFC).It is different from the traditional scheme of sending waypoints to the navigation system one by one.Firstly,the sequence waypoints are sampled in the environment map in advance.Furthermore,a virtual channel of fixed width is set by waypoints and body size.Finally,a global path is planned by the improved A*algorithm,thus realizing autonomous cruising according to the set path.The proposed scheme effectively avoids the problem of planning failure when the angle deviation between adjacent waypoints is large in the traditional approach.(2)A local path planning algorithm is proposed to improve the Timed Elastic Band(TEB)algorithm.This paper uses a PID(Proportional Integral Derivative)controller,adopts path tracking,tracks the local path obtained by the TEB algorithm,and replaces the original algorithm to calculate the output control variable,which can make the obtained control more accurate.In addition,the original TEB algorithm only constrains the maximum speed and acceleration,etc.In order to improve the performance of the algorithm in practical use,this paper adds the plus acceleration constraint to the original constraint,which limits the rate of change of acceleration,reduces the number of oscillations of the ground unmanned vehicle during obstacle avoidance,enhances the smoothness of motion.(3)By using the ROS Navigation framework and the built software and hardware simulation platform,this paper validates the hybrid path planning algorithm combining the improved A* algorithm and TEB algorithm as well as the virtual channel-based autonomous cruise scheme.Finally,the experimental results show that the proposed path planning and autonomous cruising strategies are effective and feasible in various environments.