Abstract
Although adverse effects of attacks have been acknowledged in many
cyber-physical systems, there is no system-theoretic comprehension of how a
compromised agent can leverage communication capabilities to maximize the
damage in distributed multi-agent systems. A rigorous analysis of
cyber-physical attacks enables us to increase the system awareness against
attacks and design more resilient control protocols. To this end, we will take
the role of the attacker to identify the worst effects of attacks on root nodes
and non-root nodes in a distributed control system. More specifically, we show
that a stealthy attack on root nodes can mislead the entire network to a wrong
understanding of the situation and even destabilize the synchronization
process. This will be called the internal model principle for the attacker and
will intensify the urgency of designing novel control protocols to mitigate
these types of attacks.