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.

翻译：通信延迟可能对多智能体系统造成灾难性影响。然而，现有最先进的多智能体轨迹规划器大多假设完美通信，因此缺乏在现实环境中解决这一问题的策略。为应对这一挑战，我们提出鲁棒MADER（RMADER）——一种对通信延迟鲁棒的去中心化异步多智能体轨迹规划器。通过始终保留一条保证无碰撞的轨迹并执行延迟检查步骤，RMADER即使在通信延迟下也能确保安全性。我们针对智能体间的轨迹冲突解耦进行了深入分析，开展了广泛的基准测试研究，并进行了包含多个动态障碍物的硬件飞行实验。结果表明，RMADER在无碰撞轨迹生成上实现了100%的成功率，显著优于现有方法——次优的异步去中心化方法仅达到83%的成功率。