The article introduces the Active Perception Network (APN), a novel perception roadmap for online exploration and visual sensor coverage of unknown environments. The APN serves as a hierarchical topological graph that describes how to traverse and perceive an incrementally built spatial map of the environment. It uses efficient difference-awareness techniques to track the discrete changes of the spatial map. The APN also employs a frontier-guided approach for efficient evaluation of information gain and co-visible information. The article shows that the APN outperforms several state-of-the-art approaches in terms of total exploration time and surface coverage, and is highly computationally efficient.
Publication date: 22 Sep 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2309.11695
