A multi-objective strategy based on frontier-based approach and Fisher Information Matrix for autonomous exploration

This paper presents a multi-objective exploration strategy for autonomous exploration of unknown indoor environments. The strategy mainly consists of two parts. First, it evaluates and determines the best frontier by considering the sensor information, localizability and navigation distance simultaneously. Second, a motion planning method considering the robot uncertainty is used to generate trajectories towards selected frontier. Compared to other exploration algorithms, we pay much attention to how to ensure accurate localization during the exploration and motion planning, which means the strategy should select frontiers and generate trajectories that provide sufficient information to keep the robot well-localized. Simulation experiments are presented to show the feasibility of the proposed strategy.