A greedy, deterministic sampling-based path planning algorithm for AGV

Abstract

This paper proposes a novel global path planning algorithm for Automated Guided Vehicles (AGVs). The algorithm extends the path with incremental sampling, using a greedy heuristic strategy to prioritize samples close to the goal. It also employs a vertex evaluation scheme to navigate around the obstacles. To remove redundant paths, a rewire mechanism is proposed to fine-tune the planned path. To simulate the applications in AGV, numerical simulations are conducted in $\mathbb{R}^2$ with different obstacle distributions, similar to the planar workspace of AGVs. The proposed algorithm finds better paths with less computation cost than existing open-source sampling-based planners. Compared with the optimality-guaranteed graph-search search methods, the proposed algorithm is more robust against obstacle density while ensuring a solution quality close to the global optimum.