Neuroadaptive Cooperative Fault-Tolerant Control of Heterogeneous Multiagent Systems Based on Fully Actuated System Approaches
研究了异构多智能体系统在未建模动态和执行器故障下的领导跟随协同问题，提出基于全驱动系统模型的神经自适应容错协议，使一致性误差一致最终有界，并通过数值仿真验证。
The leader-following cooperative problem in heterogeneous multiagent systems (HMASs) with unmodeled dynamics and actuator faults is investigated in this article. The HMASs, which include unmanned ground vehicles and unmanned aerial vehicles, are first described using a fully actuated system model (FASM). The FASM, as opposed to the first-order state-space model, preserves the physical significance of original systems and makes it feasible to apply the control rule entirely. In order to approximate unknown system dynamics, novel neuroadaptive laws with few learning parameters are then suggested. To counteract the negative effects of actuator faults, the Nussbaum function and adaptive approach are utilized. In addition, a cooperative fault-tolerant protocol is suggested, wherein consensus errors are uniformly ultimately bounded. The lack of virtual control variables in the proposed protocol reduces its complexity. The theoretical results are then validated by numerical simulations.