Path planning is crucial for various robots, including manipulators, that autonomously perform non-repetitive tasks in obstacle-rich environments. In this paper we focus on path planning for an n-DoF (degrees of freedom) manipulator, i.e. the one that consists of n links and n joints. We evaluate and compare head-to-head two approaches that are the most commonly used to solve this problem: search-based planning in the discretized configuration space and sampling-based planning in the continuous space. We evaluate the planners in challenging obstacle-rich scenes while varying the dimensionality of the configuration space (the number of joints/links). The memory consumption, quality of the resultant paths, runtime are tracked. Our evaluation reveals that, despite the common belief, search-based methods can outperform the sampling-based ones for 3-DoF and even 4-DoF manipulators, however for higher dimensional spaces (5-DoF and more) sampling methods are, indeed, more preferable.
DOI: 10.1007/978-3-031-77688-5_31
ResearchGate: https://www.researchgate.net/publication/387233101_Evaluating_A_and_RRT_for_High-DoF_Path_Planning
Aleksandr Onegin, Nuraddin Kerimov and Konstantin Yakovlev. Evaluating A* and RRT for High-DoF Path Planning // Proceedings of the Eighth International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’24), Volume 1. IITI 2024. Lecture Notes in Networks and Systems. Vol 1209. Pp. 324–333.