Research On The Path Planning Method Of Wheeled Robot Based On RRT

At present,mobile robot technology is one of the hot issues in scientific research,and path planning algorithm is an important research field of mobile robots.With the widening of its application scope,the development of mobile robots will becoming more intelligent and humane,which makes the accuracy of path planning more and more stringent for mobile robots.Rapidly-exploring Random Tree(RRT)algorithm based on sampling has widely used because there is no need to model the environment and has strong searchability.However,its strong randomness and slow convergence speed restrict the efficiency of mobile robot path planning.Therefore,the dissertation focused on the RRT algorithm of path planning and completes the optimization and improvement of the algorithm,then realizes the function of path planning in the wheeled robot.The main research contents are as follows:Firstly,in order to solve the problem of randomness and slow convergence rate of the basic RRT algorithm,an improved RRT algorithm is proposed by changing the selection range of key nodes and the path is optimized on the basis of variable probability guidance strategy,which verified the effect of the improved algorithm in different experimental environments.The experimental results show that the improved algorithm can accelerates the convergence speed and improves the quality of the path effectively.Secondly,in order to reduce the number of invalid nodes and improve the convergence speed of the algorithm,an improved bidirectional RRT(Bi-RRT)algorithm is proposed by setting the selection range of the new node Xnew of the extended tree and then the smallest geometric distance value Xnew' is found,then the new node will be taken as the temporary target point of the extended tree Ta to construct the extended tree Tb.The experimental results show that the improved Bi-RRT algorithm reduces the number of invalid sampling points and shortens the running time of the algorithm.Finally,the Gmapping function package of ROS system is used to construct the real environment map and configure the important parameters of the path planning.The feasibility and practicability of the improved algorithm in the actual scene are verified on the self-built wheeled robot platform.