Autonomous Exploration by Expected Information Gain from Probabilistic Occupancy Grid Mapping

Occupancy grid maps are spatial representations of environments, where the space of interest is decomposed into a number of cells that are considered either occupied or free. This paper focuses on exploring occupancy grid maps by predicting the uncertainty of the map. Based on recent improvements in computing occupancy probability, this paper presents a novel approach for selecting robot poses designed to maximize expected map information gain represented by the change in entropy. This result is simplified with several approximations to develop an algorithm suitable for real-time implementation. The predicted information gain proposed in this paper governs an effective autonomous exploration strategy when applied in conjunction with an existing motion planner to avoid obstacles, which is illustrated by numerical examples.