Exploration with active loop closing: A trade-off between exploration efficiency and map quality

摘要

A robotic system for search and rescue missions needs to efficiently explore and map new areas. In this paper, we present an integrated exploration strategy with active loop closing, which balances between the exploration speed and map quality. Specifically, it finds a trade-off between moving towards unknown space to gather new information and revisiting previous locations to improve localization accuracy and map quality through loop closures. Our integrated exploration is built upon a submap-based 6D SLAM system. Loop closure constraints originate from pairwise submap matches, which allow the optimization of an underlying SLAM graph. During exploration, we employ the expected information gain as well as the robot's localization uncertainty estimates to weigh exploration and revisiting actions online. We introduce the match effect as the expected impact of a loop closure on global optimization and consider this novel criterion together with the match likelihood and cost when evaluating the utility of revisiting previous locations. To demonstrate our approach, we present simulated and real-world experiments, comparing two variants of our novel method to a frontier-based exploration.