Event-Triggered Model-Free Optimal Consensus for Unknown Multi-agent Systems With Input Constraints

In this article, a novel event-triggered model-free structure is proposed to address the optimal consensus control problem for multi-agent systems (MASs) with unknown dynamics and input constraints. To deal with completely unknown dynamics and consensus control designs, a novel identifier-critic neural networks (NNs) structure is proposed to estimate system information and performance index function, respectively, in which novel weight adaptive laws are designed to learn unknown NNs' weights. Further, we devise an event-triggered mechanism to reduce computation loads in the framework of such a model-free optimal controller. Using Lyapunov stability theory, the stability analysis of the MASs is established. Finally, a numerical simulation is given to validate the effectiveness of the proposed control method.