Security-Driven Adaptive Iterative Learning Formation Control for Multiagent Systems
针对遭受拒绝服务攻击且状态不可测的多智能体系统，提出基于神经网络补偿和多重观测器的自适应迭代学习编队控制方案，并用仿真验证有效性。
This article addresses the adaptive iterative learning formation control (AILFC) problem of multiagent systems (MASs) with unmeasurable state subject to Denial-of-Service (DoS) attacks. To alleviate DoS attacks, a neural network (NN)-based compensation mechanism is proposed to learn communication signals, and a learning-based distributed output observer is designed to estimate the leader output. Moreover, an improved extended state observer (ESO) is designed to deal with unmeasurable states and total disturbance. Then, a time-varying boundary layer method with a normalized function is constructed to address the initial error problem. Furthermore, a multiple observer-based AILFC scheme is developed via the backstepping control technique, and the stability analysis of MASs is given by the Lyapunov theory. Finally, a simulation example is shown to illustrate the effectiveness of the developed AILFC algorithm.