Distributed fault-tolerant consensus for two-time-scale multiagent systems against multiple faults and random attacks via a generalized two-step transmission mechanism

Abstract

This paper delves into an event-based two-step transmission mechanism (TSTM) within multi-agent systems (MASs), particularly addressing the consensus control challenges posed by cyber–physical threats (CPTs). Firstly, to bolster the security and reliability of MASs in the face of CPTs, we introduce a sophisticated distributed normalized observer-controller framework which is adept at more precisely estimating unknown states and faults. Subsequently, we devised a distributed fault-tolerant consensus control (DFTCC) mechanism, which sustains the resilience of MASs against malicious attacks, compensates for system failures, and exhibits remarkable robustness to noise under challenging CPTs. Secondly, in order to mitigate network congestion, expedite data transmission rates, and optimize overall performance metrics, we propose a generalized event-based TSTM, tailored for MASs. In the initial phase, we employ a traditional event-triggered mechanism (ETM) designed to filter and temporarily store critical data trigger groups; subsequently probabilistic methods are employed to ascertain the real release packets (RRP), thereby enhancing accuracy significantly. This methodology adeptly addresses consensus challenges within MAS by substantially alleviating system burdens while ensuring instantaneous communication among components. Finally, by concurrently examining the dynamics of both fast and slow MASs through singular perturbation theory frameworks, we decompose an interrelated class of two-time-scale MASs (TTSMAS) into distinct yet discernible dynamics characterized by slower temporal scales. Moreover, through simulation experiments this methodology has proven remarkably effective in significantly enhancing the performance efficiency of MASs.