Neuroadaptive Bipartite Consensus Control for Uncertain Networked Agent Systems Subject to Actuator Attacks

Uncertainties, such as modeling inaccuracies, external disturbances, and actuator attacks are widely spread in the control channels of the networked networked agent system, which affect the accuracy and reliability of the cooperative control to a certain extent. With this in mind, this paper focuses on the robust bipartite consensus control problem of uncertain nonlinear networked networked agent systems subject to actuator attacks. In particular, the communication topology among agents is described by a signed graph, which contains not only cooperative interactions but also competitive interactions among agents. Theoretically, the neural network-based robust consensus controller is designed. On the one hand, compound uncertainties, like modeling inaccuracies and external disturbances are approximated by interpolated fitting of Gaussian radial basis functions. The adaptive laws for approximate parameters are generated online based on the stability conditions of the closed-loop error system. On the other hand, the estimation of the unknown actuator attack factor is achieved based on the adaptive control technique, which results in the effective compensation of the attack-induced control attenuation. Finally, the effectiveness of the proposed robust consensus control scheme is validated using some simulation examples.