Wang Zhenyou, Cai Xiaoquan, Luo Ao, Ma Hui, Xu Shengbing
School of Mathematics and Statistics, Guangdong University of Technology, Guangzhou 510006, China.
School of Automation, Guangdong-Hong Kong Joint Laboratory for Intelligent Decision and Cooperative Control, Guangdong University of Technology, Guangzhou 510006, China.
Sensors (Basel). 2024 Dec 11;24(24):7906. doi: 10.3390/s24247906.
This paper proposes the fixed-time prescribed performance optimal consensus control method for stochastic nonlinear multi-agent systems with sensor faults. The consensus error converges to the prescribed performance bounds in fixed-time by an improved performance function and coordinate transformation. Due to the unknown faults in sensors, the system states cannot be gained correctly; therefore, an adaptive compensation strategy is constructed based on the approximation capabilities of neural networks to solve the negative impact of sensor failures. The reinforcement-learning-based backstepping method is proposed to realize the optimal control of the system. Utilizing Lyapunov stability theory, it is shown that the designed controller enables the consensus error to converge to the prescribed performance bounds in fixed time and that all signals in the closed-loop system are bounded in probability. Finally, the simulation results prove the effectiveness of the proposed method.
本文针对具有传感器故障的随机非线性多智能体系统，提出了固定时间预设性能最优一致性控制方法。通过改进的性能函数和坐标变换，使一致性误差在固定时间内收敛到预设性能边界。由于传感器中存在未知故障，无法正确获取系统状态；因此，基于神经网络的逼近能力构建了一种自适应补偿策略，以解决传感器故障的负面影响。提出了基于强化学习的反步方法来实现系统的最优控制。利用李雅普诺夫稳定性理论，证明了所设计的控制器能使一致性误差在固定时间内收敛到预设性能边界，且闭环系统中的所有信号均以概率有界。最后，仿真结果验证了所提方法的有效性。
