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

In this paper, we study the secure consensus problem for multiple-input-multiple-output (MIMO) linear multi-agent systems (MASs) subject to denial-of-service (DoS) attacks, where an attack on an agent will block its associated communication channels until it stops. Firstly, we design an unknown input observer (UIO), based upon which we develop a resilient consensus controller, where the UIO depends only on the relative outputs. By employing a common Lyapunov function (CLF) and using the average dwell-time (ADT) method, we show that secure consensus is achieved if the attack length rate is not greater than a positive threshold. Secondly, we design a resilient consensus controller with different control gains. By using the multiple Lyapunov functions (MLFs) technique, we show that secure consensus is achieved if the attack length rate and the attack frequency are respectively not greater than the corresponding positive thresholds. Finally, we present an example of multiple YF-22 research UAVs to demonstrate the theoretical results.

Automated summary

This paper studies the problem of secure consensus for multiple-input-multiple-output (MIMO) linear multi-agent systems (MASs) under denial-of-service (DoS) attacks and proposes corresponding control schemes and conditions. By designing an unknown input observer and using multiple Lyapunov functions, it is shown that secure consensus can be achieved under certain threshold conditions.