A fault compensation strategy for consensus networks subject to transient and intermittent faults

In this paper a fault tolerant control strategy is proposed for a consensus network where some nodes can inject unexpected values as a consequence of a temporary fault. In the eventual presence of a non-collaborating node, system evolution is corrected by the action of a monitoring node. Such fault counteraction is based on an autoregressive elaboration of local data only. A perfect compensation of the drift caused by the fault on the objective function is shown to be achievable using time varying gains; the set of such time-varying gains is computed and it is shown that it satisfies a suitable autoregressive equation. Further, we estimate the convergence region of the state vector when the proposed compensation technique is used in the presence of persistent unknown but bounded disturbances. Simulations are made in order to validate theoretical results.