Research On Sampling Strategy And Path Smoothing Method In RRT Algorithm For AGV

With the concept of "Industry 4.0",the path planning technology of mobile robots has been developed by leaps and bounds in industrial production.Automated Guided Vehicle(AGV)has gradually expanded its applications in intel ligent processing plants,automatic control systems and power patrol security inspection systems due to its high flexibility and high safety.The sampling-based Rapidlyexploring Random Tree(RRT)algorithm,due to its own advantages such as no modeling,strong search ability,and the ability to search for a feasible path in a given enough time,is used in mobile robot path planning.In the field,the algorithm has been widely used and researched.This paper proposes a GO-RRT algorithm and a GOBI-RRT algorithm for AGV.The main advantage of the algorithm is that it can quickly plan a progressively optimal path that satisfies the AGV.The main content of the paper is as follows:First,a brief introduction to the definition of path planning,classification of several common path planning methods,determining the main tasks of path planning,and summarizing the design guidelines of path planning algorithms.At the same time,the advantages and disadvantages of several common path planning algorithms are analyzed in detail,and the basic RRT algorithm to be used in this article is proposed.Secondly,by analyzing the application of the basic RRT algorithm in the AGV field,there are problems such as strong randomness of its own algorithm,poor target orientation,slow convergence speed,and application of AGV without considering the size of the car body.In the selection range of sampling points,proposed to add goal-oriented ideas on the basis of variable probability guidance to optimize the path,and proposed a disc collision detection algorithm to effectively avoid the collision of the AGV body with environmental obstacles,and finally through different experimental environments Matlab in the computer verifies the effect of the GO-RRT algorithm.The results of experiment show that it effectively reduces the randomness of sampling point selection,accelerates the convergence speed of the overall algorithm,and effectively improves the quality of the pat h.Finally,by analyzing the GO-RRT algorithm,although the number of sampling points is reduced compared to the RRT algorithm,there are still many invalid sampling points,and the planned path solves the collision between the AGV body and the obstacle,but there are multiple inflection points,which make the AGV movement "detour" phenomenon.In order to further accelerate the convergence speed and remove the "detour" phenomenon of the inflection point on the AGV during the movement,the GOBI-RRT algorithm is proposed.The algorithm still adds the target guidance idea of the GO-RRT algorithm to the base BI-RRT algorithm in the basic BI-RRT algorithm,and smooth processing through the greedy algorithm on the planned path.The reverse lookup path can not be used by the node,reducing the influence of the inflection point on the path,making the path smoother.Experiments in the computer Matlab simulation and analysis multiple important parameter analysis from path plan proves that the GOBI-RRT algorithm reduces the number of invalid sampling points and inflection points,and improves the running speed of the algorithm,and the smoothed path is more suitable for AGV motion.