An enhanced Informed RRT* algorithm for path planning

Path planning is essential for enabling mobile robots to navigate in complex environments effectively. This paper proposes an enhanced Informed Rapidly-exploring Random Tree (RRT*) algorithm that reduces path planning time and generates shorter paths. The algorithm incorporates a greedy expansion strategy toward the target point, accelerating the discovery of an initial feasible path. Furthermore, a heuristic central ellipse sampling strategy refines path quality by focusing part of the sampling effort within a high-potential central region. These enhancements collectively result in faster convergence to high-quality solutions, making the algorithm suitable for complex mobile robot navigation scenarios. Compared with Informed RRT*, Python simulation results show that the proposed algorithm reduces initial path planning time by 69%−80% and path length by 9%−20%. Compared with Informed RRT*-Connect, the proposed method achieves shorter paths in the tested simulation cases, although its initial-path time is not always the lowest. Real-world experiments in two representative settings further show shorter total planning time and shorter paths.