A Distributed Algorithm for Robust Sequential Submodular Optimization in Multi-robot Systems.

With the integration of next-generation mobile communication systems and artificial intelligence, multi-robot systems utilize integrated sensing and communication (ISAC) technologies to meet the perception and communication needs for high performance, stability, and security. We introduce distributed communication strategies to enhance multi-robot cooperation, especially in sequential decision-making and complex adversarial environments. In this paper, we model these problems as a robust sequential submodular optimization framework, which restricts the multiple robots to satisfy the partition matroid constraints. We adopt the divide-and-conquer strategy to design the Distributed Robust Sequential Submodular Maximization (DRSSM) algorithm to enhance the information sharing and cooperative perception of robots. The DRSSM algorithm is proven to obtain curvature-dependent approximation guarantees. The numerical experiments demonstrate that the algorithm outperforms the centralized greedy algorithm and the centralized robust algorithm in solution quality, runtime, and robustness under attack.