Neural pinning control for adaptive trajectory tracking of complex dynamical networks

In this paper, a pinning control scheme for adaptive trajectory tracking of networked nonlinear systems is proposed. The desired trajectory tracking objective is achieved via a novel adaptive control strategy, in which the existing concept of pinning control is extended to an adaptive pinning control. The adaptive control is based on a recurrent high-order neural network identification strategy for unknown pinned nodes dynamics. The combined pinning-neural features in tracking control constitute the novelty of this approach. The proposed control law operates properly with a different reference along with unknown dynamics and coupling variations. In contrast to the linearization approach, less restrictive assumptions for node dynamics are imposed. Moreover, matrix conditions that include network topology and control properties for both pinned and non-pinned nodes are obtained. The applicability of this control scheme is illustrated via simulations.