Fast Autonomous Exploration Algorithm Based on Viewpoint Hierarchy

Autonomous exploration in complex environments is hindered by inefficient revisit planning and suboptimal local decisions. This paper presents a hierarchical viewpoint classification strategy, which incrementally identifies frontiers and samples viewpoints, to systematically coordinate global path optimization with local exploration. During routine exploration, local planners use decision goals to maximize information gain. When local goals are exhausted, a novel path deformation mechanism optimizes global revisit paths by incorporating sub-goals. Comparative evaluations in geometrically complex scenarios demonstrate that the algorithm effectively mitigates decision stagnation while maintaining exploration continuity.