Contact-Rich Collaborative Path Clearing via Physics-Informed Corridor Search

Zili Tang, Meng Guo, Zishao Qiao, Zongyuan Li

Keywords: contact-rich robotics, collaborative pushing, navigation among movable obstacles, physics-informed planning

Abstract: Contact-rich path clearing in cluttered environments requires more than detecting blocked routes: it requires selecting interventions that can reliably enlarge a width-constrained corridor through executable multi-robot physical interaction. Existing Navigation Among Movable Obstacles (NAMO) methods typically reason about obstacle rearrangement at an abstract geometric level, while collaborative pushing methods focus on local contact execution without explicitly addressing how local interactions should be chosen to produce global corridor connectivity. We present a physics-informed corridor search framework, PICHS, for contact-feasible collaborative path clearing. PICHS uses a W-clearance connectivity graph (WCCG) to determine whether a traversable corridor already exists and, when it does not, to expose the reachable bottlenecks that currently limit progress. For each selected bottleneck, the method generates only gap-conditioned pushing candidates for adjacent movable obstacles and uses short-horizon physical simulation to validate whether the induced multi-robot interactions are executable and corridor-improving. This tightly couples scene-level bottleneck selection with contact-feasible action generation, reducing unnecessary branching while preserving consistency with real physical execution. Simulation and hardware results show that PICHS improves path-clearing reliability in dense clutter and provides a practical interface between global clearance reasoning and contact-rich multi-robot control.

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: 3