摘要
Abstract
The rapidly expanding random tree algorithm(RRT)is widely used in path exploration in high-dimensional space because of its fast convergence speed and no need to model the environment.However,the application of algorithms in complex envi-ronments also exposes some problems,such as the algorithm's blind generation of random points leads to too long search time,re-dundant path generation,unsmooth,and so on.To solve the above problems,an improvement scheme is proposed.First,the con-cept of a preset tree is proposed.The exploration area is evenly distributed to the entire space.At the same time,the RRT connect search method is introduced.The target bias factor,greedy strategy,and variable step rectangle scheme are added to make the dou-ble trees connect as quickly as possible.Then,after using the global optimization strategy,the forward and reverse pruning are used to obtain a better set of advantages.Finally,the nurbs curve is used to smooth the path interpolation.The simulation experiment shows that the search time of the iterative algorithm RRT*in this scheme is reduced by about 76.5%,the path length is reduced by about 11%,and the time to find the path of the search algorithm RRT connect is shortened by about 55%.关键词
RRT算法/预置树/曲线平滑/最优路径Key words
RRT algorithm/preset tree/smooth curve/optimal path分类
计算机与自动化
