Data-Driven Adaptive PID Control for Unknown Nonlinear MAS with Input Constraint

This paper addresses the consensus tracking problem for a class of unknown nonlinear multi-agent systems (MAS) subject to input constraints. We present a new data-driven adaptive PID control method. This approach allows for decentralized coordination without needing clear system dynamics. Based on the dynamic linearization technique, a linear data model that transforming over time is first constructed for the nonlinear system. Subsequently, by integrating the PID control framework, a data-driven control algorithm was developed that relies solely on the saturated input data and output data of the follower agents and their neighboring nodes. Additionally, the contraction mapping principle and mathematical induction method are used to formally show the proposed algorithm’s convergence under two scenarios. Lastly, simulations are used to confirm the suggested algorithm’s viability and efficacy.