An Efficient Beam Search Algorithm for Active Perception in Mobile Robotics

Kaixian Qu, Han Wang, Victor Klemm, Cesar Cadena, Marco Hutter

Keywords: active perception, planning algorithms, mobile robotics

Abstract: Active perception is a fundamental problem in autonomous robotics, in which the robot must decide where to move and what to sense in order to obtain the most informative observations for accomplishing its mission. Existing approaches often solve a computationally expensive traveling salesman problem over heuristically selected nodes, or adopt an overly constrained shortest path tree formulation. We introduce node-wise beam search (NBS), an efficient algorithm that maintains the top candidate paths per node to effectively explore the solution space. To balance exploration and exploitation, we integrate frontier information into a novel expected gain metric. Furthermore, we propose the rapidly-exploring random annulus graph (RRAG), an incremental graph construction method that preserves full orientation sampling and ensures connectivity in cluttered environments via a fallback local planner. Evaluations across three active perception tasks demonstrate that NBS combined with RRAG outperforms state-of-the-art baselines by at least 20\% in one or more tasks.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 4