Studies on sampling based path planning algorithms in autonomous mobile robot navigation

Abstract

Mobile robotics is one of the fastest-growing fields of science in the newlinemodern era. Due to their capabilities, autonomous mobile robots are widely newlineused in the industrial environment for warehouse material handling and newlinelogistics, in the radioactive environment for nuclear decommissioning and newlineradioactive material search, in agriculture for automated precision agriculture newlineoperations, in space exploration, and the hospital environment for hospital newlinedisinfection and transportation. To accomplish these tasks, the autonomous newline-planning algorithm must be efficient. By avoiding newlineobstacles in its way, the path planning algorithm finds a possible path from newlinethe start point to the end point. newlineSampling-based planning algorithms are widely used in robotics newlinedue to their low procedure computational cost for high dimensional issues and newlineprobabilistic completeness. The most popular sampling-based path planning newlinealgorithms are Probabilistic Roadmap (PRM) and Rapidly Exploring Random newlineTree (RRT). Even though RRT is faster than the PRM for single-query newlineproblems, the path found by RRT is not optimal. An extension to RRT, called newlineRRT*, provides an asymptotically optimal solution. The sampling scheme is newlinethe heart of the RRT* path planning algorithm. It is classified into uniform newlinesampling and non-uniform sampling, also known as informed sampling. newlineHowever, RRT* suffers from a slow convergence problem due to the uniform newlinesampling of the entire configuration space. Hence, the non-uniform sampling newlineprocess in RRT* improves the convergence of RRT* by limiting the search newline