Performance Guaranteed Composite Adaptive Neural Consensus Control for Uncertain Multiagent Systems

This article studies performance guaranteed composite adaptive neural consensus control problem of uncertain multiagent systems over a directed graph. First, a piecewise function is adopted to characterize fixed-time prescribed performance (FPP), and barrier functions are constructed to achieve the prescribed system performance. Next, hybrid echo state networks (HESNs) are constructed and employed to approximate unknown nonlinear dynamics. To realize smooth and fast learning, adaptive neural predictors are designed, and the predictor errors are fused with barrier functions or virtual control errors to design composite adaptive weights update laws of HESNs. Then, a novel FPP composite adaptive consensus control strategy is proposed. It ensures prescribed tracking performance and high-precision NN approximation performance. Finally, the availability of the proposed strategy is confirmed through simulation studies.