Improved networked predictive control with different transmission delays in both forward and feedback channels

The paper is concerned with an improved predictive control method of NCSs (Networked Control Systems) with random time delay and data dropout. The design of networked control system we proposed is based on Scenario-Based Networked Control (SBNC). Through this method the NCSs' performance is relative with an equivalent data-dropout probability. In this paper, the improved method is applied on the linear systems, and the network is in both forward and feedback channels that is more realistic. The buffer in feedback channel makes the system equivalent to a system with an extra forward delay, and the feedback channel has no delay. The simulation on the ball and beam system verifies the effectiveness and rationality of the proposed control algorithm. Besides, the comparison between a different predictive method and this improved one is achieved, and the latter one shows better tracking properties even when the network is very bad. The effects of the parameters are also considered in the simulation.