Step adaptive controller for networked MIMO control systems

This paper considers controller tuning for multivariable processes in networked control systems (NCSs) and presents a novel control strategy. In MIMO NCS systems the controller has to manage both the cross-interactions between different loops, and the delays and losses induced by the network. We investigate the effect of cost function weighting and cross-interaction restriction in MIMO systems by simulation-based optimization of controller parameters. Based on this, a step adaptive controller is developed, which changes its tuning parameters depending on whether step response or cross-interaction elimination is desired. We also analyze the proposed control scheme from the networked control perspective via the jitter margin. The step adaptive controller is compared with other MIMO control methods, both in a traditional MIMO case, but also in a NCS setup with the communication and control co-simulator PiccSIM. The results show that the step adaptive controller performs well in comparison to methods presented previously in the literature, and it can suppress the cross-interaction efficiently.