Delay Distribution-Based Smith Predictor for Networked Control Systems

In this work, we study the discrete-time networked control systems (NCSs) with network-induced dual channel delays. In a sampling period, if the feedback-channel delay is historical information for the controller node, then the forward-channel delay is an unknown future quantity. The typical Smith predictor scheme usually treats the forward-channel delay as a measurable historical quantity, which makes it not reasonably compensated. In this paper, a novel Smith predictor scheme is proposed to compensate for dual-channel delays and assist in controller design, which utilizes information about the distribution of forward-channel delays to predictively compensate for them. Then, the methods for determining the stability condition of the system and the controller feedback gain that account for delay distribution are provided. Finally, two simulation examples are given to show that the proposed scheme achieves better state convergence and tracking performance under varying network conditions compared to typical Smith predictor schemes.