Accurate mapping is indispensable for reliable indoor robot navigation. Active Simultaneous Localization and Mapping (Active SLAM) enables a robot to localize itself while actively exploring unknown environments to construct maps. This paper presents a Traveling Salesman Problem (TSP)-based complete active SLAM framework designed for efficient map generation. The proposed method generates grid-based waypoints within polygonal regions of interest, formulates the problem as a TSP, and applies a nearest-neighbor heuristic to compute a traversal path starting from the waypoint closest to the robot’s current position. The robot follows the planned trajectory, incrementally building the map and avoiding obstacles, and returns to the starting position after visiting all waypoints, thus completing active SLAM. The approach was validated on an indoor robot implemented in the Robot Operating System (ROS). Experimental results demonstrate that the proposed algorithm achieves accurate mapping and reliable navigation. Comparative evaluations against Active Graph SLAM and RRT-SLAM in the same environment confirm that our method attains comparable mapping accuracy while significantly reducing the time required for map completion.
Kim et al. (Thu,) studied this question.