Distributed impulsive control for secure consensus of nonlinear dynamical networks with node delay under denial‐of‐service attacks

Abstract

This paper investigates secure consensus problems of nonlinear dynamical networks (NDNs) with node delay under denial‐of‐service (DoS) attacks, which can cause a break in communication links. It will lead to packet loss. To ensure the secure consensus in this scenario, we propose an effective impulsive control which only based on sampled information from neighbor nodes. Based on the Lyapunov stability theory and the impulsive comparison principle, an upper bound of the average impulse interval is derived to ensure the exponential consensus of NDNs under DoS attacks, and the convergence rate can be estimated. Finally, a numerical example is provided to check the theory results.

Secure consensus of multi‐agent systems under denial‐of‐service attacks

“…(1) Compared to existing works [27][28][29][30][31][32][33][34], in which the resilient controllers were developed based on (relative) full states, the designed resilient controller here depends only on relative outputs and therefore are more practical for real applications. (2) The DoS sequence is not required to follow any stochastic distribution as required in previous studies [27,29].…”

“…Compared to existing works [27–34], in which the resilient controllers were developed based on (relative) full states , the designed resilient controller here depends only on relative outputs and therefore are more practical for real applications.…”

“…In previous works [1,[14][15][16][17][18][19][20][21][22], the controllers are designed under secure communication environment, which however is unrealistic due to ubiquitous cyber attacks [24,25]. Actually, how to achieve secure consensus in MASs under insecure communication environment has already raised great concerns [26][27][28][29][30][31][32][33][34]. For MASs subject to random cyber attacks, in Feng et al [27], a hybrid stochastic secure control framework is established for achieving mean-square consensus.…”

“…The contributions of this paper are listed as follows: Compared to existing works [27–34], in which the resilient controllers were developed based on (relative) full states , the designed resilient controller here depends only on relative outputs and therefore are more practical for real applications. The DoS sequence is not required to follow any stochastic distribution as required in previous studies [27, 29]. This paper introduces the attack length rate and the attack frequency to characterize the attack sequence which is easier to verify. The UIO technique was used in previous studies [14, 21, 22] to solve the consensus problem for MASs under secure communication environments, but this paper investigates the more challenging case under insecure communication environments. …”

Performance limits of network delay systems based on bandwidth and packet loss constraints

“…In [14], the authors present a detailed work which focuses on the stability of NCSs on the basis of network contained packet loss and delay. The work presented in [15] models the NCSs based on network-induced delay. In addition, the stability of the control system and the design of the controller are also studied.…”