DiscoverBot: Advancing Autonomous Exploration through Frontier and Informed-RRT* Fusion in Mobile Robots

摘要

Exploration is a fundamental task within the autonomous robotics domain, ranging from search and rescue robotics to environmental monitoring and space domains. This paper presents a new frontier-based exploration approach, using the clustering of points with DBSCAN, path planning by Dubins Informed-RRT* , and path following with a pure pursuit controller on a TurtleBot. The autonomous navigation process will be further enhanced in this way with respect to the optimization of the exploration process and path optimization. To begin with, frontier points are identified and clustered using the DBSCAN algorithm in order to get the best target viewpoints. Then, the Informed-RRT* algorithm is used in the development of the path planning algorithm with the Dubins Path. It is capable of generating paths that are shorter than RRT. This is gained by its capacity to search for heuristic-generated ellipsoids that can be used to eliminate the computational overhead. Finally, the path is followed by the mobile robot with a pure pursuit controller. On the simulation and real-world experiments of the proposed solution, a significant improvement in exploration efficiency and path quality is shown. The Informed-RRT* finds out the effectiveness of the algorithm with respect to path optimality and computational efficiency over RRT*. This paper further improves autonomous navigation by providing a highly robust and efficient framework for exploration and path planning.