Comparison of Path Planning Algorithms for Indoor Static Environment

Indoor path planning application area involves cleaning, surveying, inspection & mobile robotics etc. In this paper, we performed a comparative analysis of number of well-known algorithms for indoor environments, to find the best algorithm providing the shortest path & execution time. The obstacle space was simulated using a 2D environment created using a binary grid map of 200x200 cells, which represents indoor environment features like walls, narrow passages, closed rooms, door openings, etc. Popular methods for path planning include visibility graphs, grid-based methods, sampling-based methods, virtual potential fields, etc. Each of these methods is suitable for a certain environment. Complete path planners like visibility graphs are mathematically complex & impractical for many real systems. Grid-based planners like A* or wavefront planners are easy to implement but are suitable for low-dimensional configuration space. Planners that rely on sampling such as Rapidly-exploring random tree (RRT) & Probabilistic roadmap (PRM) are computationally less expensive than grid-based planners & rely on a random function to choose a sample from configuration space to find obstacle-free space & a simple local planner to connect these samples to move towards a goal. We have selected a variety of algorithms, each representing different path planning criteria involving simple bug behavior, sampling-based planners, distance cost function with a heuristic approach & compared the final results to find a suitable algorithm for the indoor environment. It was observed that sampling based & heuristic-based distance function planner performs best among selected algorithms for the indoor environment under consideration.