Enhancing Path Planning for Autonomous Robots in Large, Obstacle-Crowded Environments: A Practical Improvement to the PRM Algorithm

Probabilistic roadmap (PRM) approximation algorithm has been successful in solving many motion planning problems. However, when faced with crowded areas, PRM tends to suffer from excessive computation times and produce suboptimal solutions. In this research, we present an enhancement to PRM and introduce a novel PRM-type algorithm that offers notable improvements in computation time, convergence speed, and path length compared to its classical counterpart. The key innovation of our algorithm lies in its adaptability to dense environments, achieved by checking several most suitable directions and choosing the least crowded one. Additionally, when encountering obstacles, the algorithm searches for detour options in relatively small, obstacle-crowded subareas rather than processing each obstacle individually or the entire map. This allows self-navigation robots to adjust planned paths in real-time, ensuring smoother, quicker, and obstacle-free routes. Experimental results demonstrate that our approach consistently reduces computation time, final route length, and the number of nodes compared to various PRM variants. Specifically, across different obstacle and node densities, the proposed algorithm outperforms the PRM benchmark, confirming its effectiveness for path planning in both simulated and real-world environments, particularly within large-sized areas.