Event-Triggered Data-Driven Formation Control for Heterogeneous Multiagent Systems Subject to DoS Attacks

This article studies the secure robust formation control problem for heterogeneous multiagent systems involving unmanned air–ground vehicles subject to denial-of-service (DoS) attacks, external disturbances, and unknown system parametric information. For the heterogeneous agents involving nonlinear underactuated coupled dynamics and limited communication capabilities, a unified event-triggered data-driven control method is proposed, including a distributed event-triggered observer for interaction data recovery under DoS attacks and robust controllers to achieve distributed formation. The control policies and disturbances, including couplings and external disturbances, are modeled as performance optimizers and adversarial players in the zero-sum game. Robust optimal policies are obtained by developing off-policy reinforcement learning without information on system dynamical parameters. Sufficient conditions are derived to guarantee the stability of the closed-loop system under DoS attacks, and Zeno behavior for all agents can be excluded. The effectiveness of the proposed controllers is validated by presenting two simulation examples of a distributed heterogeneous formation.