Data-Driven Output Consensus Tracking Control for Heterogeneous Multi-Agent Systems with a Dynamic Leader

This paper investigates the output consensus tracking problem for a class of heterogeneous multi-agent systems with a dynamic leader, where the followers' system dynamics are completely unknown. In this context, the leader's control input is nonzero, time-varying, and inaccessible to all followers. To tackle such a problem, the model-based distributed control scheme is developed via distributed continuous observer design. Subsequently, the data-driven distributed control scheme is proposed, which comprises a compensation term, an optimal feedback term, and a feedforward term. Specifically, a new kind of distributed adaptive continuous observers is first constructed based on the adaptive control theory to estimate the leader's state, output, and plant matrices, the off-policy learning method is then employed to seek the optimal control gain and identify the unknown system dynamics, thereby facilitating the construction of the control protocols. It is proved that the multi-agent system driven by the designed control laws can achieve output consensus tracking. Finally, the theoretical results are validated through performing simulation experiments.