LESE-GAE: Lightweight Environment Skeletonization and Ellipse-Heuristic Goal-Motivated Autonomous Exploration

Autonomous exploration has extensive applications and demonstrates significant potential for navigating in unknown environments. However, achieving high-efficiency goal-motivated autonomous exploration with limited hardware resources still remains challenging. In this work, we propose a lightweight environment skeletonization and ellipse-heuristic goal-motivated exploration framework for fast autonomous exploration. A high-efficiency perception region reconstruction and a polar-based frontier detection algorithm are proposed to directly generate topological maps from the raw sensory data, supporting the incremental construction of the lightweight environmental skeleton. Built upon the skeleton, an ellipse-heuristic exploration strategy is proposed, where ellipses are constructed with positions of the goal and the robot, and the corresponding eccentricity and major axis length are utilized to evaluate the global guidance capability of the detected frontiers efficiently. Both simulation and real-world experiments show that, compared with the state-of-the-art methods, our method can reduce storage consumption and computation time for map updating by over 83.84% and 95.70%, respectively. Furthermore, it is the only method that successfully completes all exploration tests, achieving over 19.71% reduction in exploration time.