Attack-Resilient Distributed Optimization-based Control of Multi-Agent Systems with Dual Interaction Networks

The vision for the Urban Air Mobility airspace is a highly automated, cooperative, passenger and/or cargo carrying air transportation service for economic purposes. Comprised of highly complex, safety critical cyber-physical systems (CPSs), the integration of a UAM system within the National Airspace System (NAS) requires the development of robust control paradigms that are resilient to cyberattacks. Consequently, the cybersecurity of CPSs has emerged as one of the most important issues for UAM general operations. In this paper, we consider Denial-of-Service (DoS) cyberattacks and their effects on UAM traffic synchronization and collision avoidance for agents (e.g., aerial vehicles) within the modeled UAM flight corridor. Network connectivity is essential for such models to perform tasks such as distributed optimal control, optimal consensus, or distributed optimization problems, in a collaborative manner. We propose a novel distributed optimization-based control strategy that prompts UAM vehicles, who self-identify as vulnerable, to move towards the centroid of the network to maintain connectivity. We construct a composite interaction network by mixing the redundant information from the communication and auxiliary sensing networks to robustify the communication edge links so that the UAM vehicles are less likely to be vulnerable in the event of a DoS cyberattack. We validate the performance of the proposed control strategy via an illustrative simulation for aerial vehicle traffic synchronization and collision avoidance in the UAM flight corridor, in the presence of DoS cyberattacks.