Exponential consensus of stochastic discrete multi-agent systems under DoS attacks via periodically intermittent control: An impulsive framework

Highlights：

• For SDMASs without DoS attacks, we derive the relation among the control period, the control width and the impulsive interval via utilizing the Lyapunov method and the linear matrix inequality technique. In addition, for SDMASs with DoS attacks, we derive the control rate associated with the impulsive interval, the impulsive gain, the DoS attack ratio and the attack intensity under reasonable assumption that adversaries fail to work persistently within given intervals.

• Motivated by the notion of average impulsive interval, a novel concept of periodic window-based average impulsive interval is presented to make our developed results less conservative, which can avoid employing the maximum and minimum impulsive intervals.

摘要

•Integrating the advantages of intermittent control and impulsive control, a novel periodically intermittent impulsive control (PIIC) scheme is proposed to address the exponential leader-following consensus problem of stochastic discrete multi-agent systems (SDMASs) with and without DoS attacks, where the impulsive effect is available at the control windows only. In addition, comparing with the previous works, the discrete-time dynamics studied in this article is more comprehensive, which takes into account nonlinearities, parameter uncertainties, stochastic disturbances and DoS attacks.•For SDMASs without DoS attacks, we derive the relation among the control period, the control width and the impulsive interval via utilizing the Lyapunov method and the linear matrix inequality technique. In addition, for SDMASs with DoS attacks, we derive the control rate associated with the impulsive interval, the impulsive gain, the DoS attack ratio and the attack intensity under reasonable assumption that adversaries fail to work persistently within given intervals.•Motivated by the notion of average impulsive interval, a novel concept of periodic window-based average impulsive interval is presented to make our developed results less conservative, which can avoid employing the maximum and minimum impulsive intervals.