Model-Free Adaptive Control of Second-Order Multi-Agent Systems via Backstepping Under Mixed Attacks

This paper employs model-free adaptive control methods to investigate second-order multi-agent systems under mixed attacks. Based on backstepping techniques, second-order multi-agent systems are equivalently transformed into two interconnected first-order subsystems to reduce structural complexity. Additionally, considering the impacts of multi-source mixed network attacks,compensation mechanism employing predictive distributed output are designed, which mitigate the negative consequences of such attacks. Furthermore, leveraging only input-output data from each subsystem, a distributed modelfree adaptive control scheme utilizing backstepping is developed to achieve consensus. Finally, the effectiveness of the proposed method is validated through two simulation examples.