Safe SLAM exploration strategy using optimal path planning towards frontier points
Nikolaos Evangeliou, Omar Mostafa, Anthony P. Tzes · 2025
A novel safe exploration strategy for Simultaneous Localization and Mapping (SLAM) is presented. The safety in the strategy is based on the utilization of a Medial Axis planner for generating optimal paths through a combination of fixed obstacles and frontier points at the intersection of explored and unexplored areas of a map. The current approach is tailored for Autonomous Ground Vehicles (AGVs), where a 2D exploration strategy needs be deployed as part of the SLAM pipeline. Inhere, an omniwheel AGV is deployed, capable of performing SLAM using the RTAB-Map SLAM pipeline, using a combination of accelerometers, gyroscopes, a 3D-LiDAR and a forward-facing RGB camera. The developed algorithm leverages a safe exploration approach at the dichotomous of available headings for known global map and a fallback routine aiming to push the exploration to frontier points in an optimal obstacle avoidance manner. Experimental studies with the mobile agent are presented indicating the efficacy of the proposed strategy.