Robust adaptive consensus of nonstrict-feedback multi-agent systems with quantized input and unmodeled dynamics

Abstract

This paper deals with the distributed robust adaptive neural network tracking control for a class of nonstrict-feedback multi-agent systems (MASs) with unmodeled dynamics and input quantized by the hysteresis quantizer. A decomposition technique is utilized to handle the problem that only finite and discrete control values are available to control system. Meanwhile, the radial basis function neural networks (RBFNNs) are employed for approximation of the packaged unknown nonlinearities induced by recursive design. By introducing the structural characteristic of Gaussian basis function (see Lemma 5), the effects from the nonstrict-feedback are effectively compensated and the backstepping-based method can be successfully implemented. Then, by combining the dynamic signal with adaptive backstepping control technique, a robust adaptive distributed quantized tracking control scheme is developed for each follower. It is proved that all closed-loop signals are semi-globally uniformly ultimately bounded and the asymptotic consensus tracking is guaranteed. Finally, the effectiveness of the control method is illustrated by two simulation examples.