Resilient Consensus for Incremental Quadratic Constrained Nonlinear Multi‐Agent Systems Under DoS Attacks and Unmatched Unknown Inputs
TL;DR
Simulation results validate the effectiveness of the proposed approach in maintaining consensus performance despite the presence of non‐vanishing disturbances and intermittent DoS attacks.
Abstract
This paper addresses the problem of resilient consensus for a class of nonlinear multi‐agent systems (MASs) subject to denial‐of‐service (DoS) attacks and unmatched unknown inputs. A significant challenge in such systems arises from the simultaneous presence of strong Lurie‐type nonlinearities and disturbances that are unmatched with the control inputs, necessitating explicit reconstruction via an algebraic mechanism. To overcome these limitations, a robust Circle‐Criterion unknown input observer (UIO) equipped with a nonlinear injection term is proposed. By employing an auxiliary output construction technique, the strict rank condition typically required for disturbance decoupling is relaxed, thereby enabling effective state estimation and algebraic reconstruction of unknown inputs even in rank‐deficient scenarios. Furthermore, to mitigate the impact of DoS attacks, a dual‐mode resilient control protocol is developed, which switches between observer‐based feedback during normal communication and an open‐loop predictor during attack intervals. Rigorous stability analysis is conducted utilizing input‐to‐state stability (ISS) properties and the average dwell time (ADT) method. Sufficient conditions for achieving bounded consensus are derived in the form of linear matrix inequalities (LMIs). Simulation results validate the effectiveness of the proposed approach in maintaining consensus performance despite the presence of non‐vanishing disturbances and intermittent DoS attacks.