Output Tracking Predictive Control of Networked Systems with Two-channel Random Communication Constraints

In this paper, the output tracking control problem is investigated for a networked control system with two-channel random network delays and packet dropouts as well as stochastic noise. To actively compensate for these random communication constraints in the feedback and forward channels, a novel networked predictive control (NPC) method is proposed based on the input-output difference equation model, where the two-channel communication constraints are handled separately according to their different features. Furthermore, different from the existing NPC methods based on round-trip time delays, actual control inputs rather than the predicted ones are employed to generate future control commands. Then a delay-independent closed-loop stability condition is obtained, and a condition to guarantee a zero steady-state output tracking error is derived. Also, theoretical analysis shows that the proposed NPC method can achieve the same output tracking performance as the corresponding local control system. Finally, the effectiveness of the proposed method is evaluated by simulation and experimental results.