Synchronized Multi-arm Rearrangement Guided by Mode Graphs with Capacity Constraints

Solving task planning problems involving multiple objects and robotic arms poses scalability challenges. Such involve not only coordinating high-DoF arms, but also searching through possible sequences of actions including object placements, handoffs. The current work identifies a useful connection between multi-arm rearrangement recent results in multi-body path on graphs with vertex capacity constraints. synchronized at high-level involves reasoning over modal graph, where nodes correspond to stable placements transfer states by the arms. Edges this graph pick, placement handoff operations. can be viewed as pebbles moving which has For instance, each arm carry single locations accumulate many objects. Efficient integer linear programming-based solvers have been proposed for corresponding pebble problem. proposes heuristic guide process rearrangement. Results indicate good objects, an algorithm that find high-quality solutions fast exhibiting desirable anytime behavior.