A Unified Approach for Autonomous Volumetric Exploration of Large Scale Environments Under Severe Odometry Drift

Exploration is a fundamental problem in robot autonomy. A major limitation, however, that during exploration robots oftentimes have to rely on on-board systems alone for state estimation, accumulating significant drift over time large environments. Drift can be detrimental safety and performance. In this work, submap-based, multi-layer approach both mapping planning proposed enable safe efficient volumetric of scale environments despite odometry drift. The central idea our combines local (temporally spatially) global guarantee efficiency. Similarly, leverages the presented map compute frontiers changing utilizes nature dealing with partial information planning. system thoroughly evaluated shown outperform art methods even under drift-free conditions. Our system, termed GLoca}, will made available open source.