Observer-based data-driven consensus control for nonlinear multiagent systems with faded neighborhood information

A data-driven consensus control approach is proposed for unknown nonlinear multiagent systems (MASs) with faded neighborhood information and disturbances under switching topologies. Since the exact model of MASs is unknown, each agent is given an equivalent affine data model using a local dynamic linearization method. To estimate the time-varying parameter in the affine data model, an adaptive update algorithm is designed. The total uncertainty induced by parameter estimation error and unknown nonlinearity is estimated using a discrete extended state observer and compensated in the control law. Further, considering the impact of the faded neighborhood information, a compensation term is introduced to correct the output information. Then, a data-driven consensus control protocol is proposed, and the asymptotic convergence of the tracking error is proven. Finally, the simulation study verifies the theoretical results.