Robust MADER: Decentralized Multiagent Traj Planner Robust to Comm Delay in Dynamic Environments
Accepted to IEEE Robotics and Automation Letters (RA-L)
arXiv: http://arxiv.org/abs/2303.06222
Code: https://github.com/mit-acl/rmader
ResearchGate: https://www.researchgate.net/publicat...
Abstract— Communication delays can be catastrophic for multiagent systems. However, most existing state-of-the-art multiagent trajectory planners assume perfect communication and therefore lack a strategy to rectify this issue in real-world environments. To address this challenge, we propose Robust MADER (RMADER), a decentralized, asynchronous multiagent trajectory planner robust to communication delay. By always keeping a guaranteed collision-free trajectory and performing a delay check step, RMADER is able to guarantee safety even under communication delay. We perform an in-depth analysis of trajectory deconfliction among agents, extensive benchmark studies, and hardware flight experiments with multiple dynamic obstacles. We show that RMADER outperforms existing approaches by achieving a 100% success rate of collision-free trajectory generation, whereas the next best asynchronous decentralized method only achieves 83% success.
![[ICRA24] PUMA: Decentr. Uncertainty-aware Multiagent Traj. Planner w/ Image Segmentation Frame Align](https://images.videosashka.com/watch/W73p42XRcaQ)
![[Presentation] Deep-PANTHER: Learning-Based Perception-Aware Traj. Planner in Dynamic Environments](https://images.videosashka.com/watch/zlMxuSSvo3o)
