This paper introduces a distributed adaptive fault-tolerant control scheme for nonlinear multi-agent systems afflicted by actuator failures. The proposed scheme ensures that the consensus-tracking errors of the systems meet the prescribed performance requirements. First, in the backstepping design process, a barrier function is constructed based on the proposed scalar function and normalization function to address the prescribed performance issue. Next, both multi-dimensional Taylor network (MTN) control and robust control techniques are utilized to address unknown nonlinear functions. Continuous switching functions are devised to transition from MTN control to robust control whenever the arguments of the unknown functions surpass the operating range of MTN. This effectively resolves the global problem of the system. Furthermore, a boundary estimation method is proposed to handle actuator failures that may occur infinitely. Stability analysis confirms that all signals in the closed-loop system are globally uniformly ultimately bounded. Additionally, the consensus-tracking errors can always be constrained within prescribed boundaries, achieving global convergence to a prescribed accuracy within a prescribed time. Users can flexibly and independently choose the desired settling time and accuracy, without being restricted by any initial conditions. Simulation results provide evidence of the efficacy of the proposed control method.

中文翻译：

---

非线性多智能体系统具有规定性能的分布式自适应容错控制


本文介绍了一种针对受执行器故障影响的非线性多智能体系统的分布式自适应容错控制方案。所提出的方案确保系统的共识跟踪误差满足规定的性能要求。首先，在反步设计过程中，基于所提出的标量函数和归一化函数构造障碍函数，以解决规定的性能问题。接下来，利用多维泰勒网络（MTN）控制和鲁棒控制技术来解决未知的非线性函数。当未知函数的参数超过 MTN 的工作范围时，连续切换函数被设计为从 MTN 控制过渡到鲁棒控制。这有效地解决了系统的全局问题。此外，提出了一种边界估计方法来处理可能无限发生的执行器故障。稳定性分析证实闭环系统中的所有信号都是全局一致最终有界的。此外，共识跟踪误差始终可以被限制在规定的边界内，从而在规定的时间内实现全局收敛到规定的精度。用户可以灵活、独立地选择所需的稳定时间和精度，而不受任何初始条件的限制。仿真结果证明了所提出的控制方法的有效性。