Fast finite-time adaptive optimal algorithm for stochastic multiagent systems under deception attacks and time-varying output constraints

Abstract

An adaptive optimal neural network algorithm with fast finite-time convergence is proposed for stochastic multiagent systems (SMASs) under deception attacks, time-varying asymmetric output constraints and dead zones. An additional attack signal corrupts the state information of nonlinear systems, resulting in the unavailability of real state information for controller development. To overcome this obstacle, a reinforcement learning (RL)-based identifier-actor-critic-disturbance architecture is used to develop a fast finite-time adaptive optimal tracking algorithm for each subsystem in SMASs, which alleviates the negative effects of cyberattacks that intentionally tamper with sensor signals. Herein, a barrier function is designed to transform the constrained system into an unconstrained equivalent. Furthermore, time-varying dead zones in SMASs pose considerable challenges for controller design, while enhancing the applicability of the system in practical scenarios. The proposed resilient adaptive optimal tracking algorithm guarantees the boundedness of all signals in the overall system in probability. Eventually, two simulation results are conducted to prove the effectiveness of the proposed method.