Belief-Aware Multi-Agent Path Finding under Map Uncertainty
MAGIC updates shared traversability beliefs so multi-agent pathfinding avoids costly reroutes under unknown map changes.
Classical multi-agent path finding assumes known obstacles, but real maps change. MAGIC uses a Gaussian Markov Random Field and Gaussian Belief Propagation to infer traversability from spatially correlated observations and build detour-aware costs. On MAPF benchmarks it reduced executed sum of costs versus existing methods on 96.3% of instances, including teams of up to 800 agents.
- Map discrepancies stay fixed but start unknown.
- Gaussian belief propagation infers nearby unobserved obstacles.
- MAGIC beat baselines on 96.3% of instances.
- Tested across planner families with up to 800 agents.
From the original article
1 min read at arxiv.orgMulti-Agent Path Finding (MAPF) aims to find collision-free paths for multiple agents in a shared environment. Classical MAPF assumes that all static obstacles are known in advance, but real-world environments can change unexpectedly due to fallen objects, spills, or other local disturbances. When such changes are spatially correlated, an observation can inform traversability estimates beyond the observed location.
By Viraj Parimi · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.