Distributed Fault Tolerant Control of Multi-Agent System Based on LSTM Trajectory Prediction Model Under Intermittent Communication Faults

This paper proposes a novel trajectory prediction based fault-tolerant control method for multi-agent systems under intermittent communication faults. Agents affected by intermittent communication faults are unable to communicate with neighbors or obtain reference tracking information for a period of time. To address this issue, we first employ a trajectory prediction model trained using long short term memory neural network to predict the expected trajectory in the event of intermittent communication faults. Subsequently, a distributed controller based on fixed-time theory is designed to track the predicted values. Compared with the existing fault-tolerant control methods, the proposed strategy achieves smaller consensus errors and a longer tolerable intermittent inter-ruption interval. Moreover, it demonstrates the capability to predict the actual expected trajectory trends of agents during large-angle maneuvers. Finally, the effectiveness of the fault-tolerant control strategy is verified by simulation.