Segmented Approach to Path Planning

Abstract

A* algorithm performs well as a Best First Search method, which would not give the shortest path in certain scenarios. Its accuracy depends on the heuristic function and has slow processing speed in the real world. RRT performs slower than A* and Dijkstra's algorithm gives correct output but shows us a slow runtime performance unsuitable for the real-world. This paper uses Dijkstra's algorithm using the priority queue for testing and proposes an approach that can be applied to any path planning algorithm. Experimental results show that the proposed approach performs 51% faster than A* on game datasets and 14% faster on extremely dense map datasets.