Distributed fault-tolerant for leader-following multi-unmanned aerial vehicle systems with faulty sensors based on belief rule base

Effective fault-tolerant control (FTC) for leader-following Multi-Unmanned Aerial Vehicle (Multi-UAV) systems is crucial but challenging, especially when dealing with sensor faults in a distributed manner, including complex scenarios like overlapping failures leading to joint unobservability across multiple UAVs. This paper proposes a novel distributed FTC framework leveraging Belief Rule Base (BRB) combined with communication fusion (BRB-CF) to address these challenges. The framework integrates two key modules operating on each UAV: a decentralized BRB-based fault diagnosis (FD) module to accurately identify faulty sensor locations using local data, and a distributed reconstruction module utilizing BRB-CF. For non-overlapping faults, a communication-based fusion function compensates failures using neighbor data. Crucially, for overlapping faults where fusion is insufficient, a dedicated BRB-CF model reconstructs sensor outputs by integrating historical data with fused neighbor information. Simulation results demonstrate the proposed framework's effectiveness, enabling the multi-UAV system to maintain formation and performance stability under both non-overlapping and overlapping sensor fault conditions, thereby enhancing system robustness and resilience. © 2025 Elsevier Ltd.