Decentralized predictive control for tracking constant references

This paper describes a novel Decentralized Predictive Control (DePC) technique for tracking constant reference signals. The controlled system is supposed to be constituted by a set of non-overlapping subsystems coupled by states and inputs. First the offset-free tracking problem is recast as a regulation one by reformulating the plant model in the so-called “velocity-form”; secondly the decentralized control problem is solved by resorting to “tube-based” robust MPC, where dynamic interactions between subsystems are interpreted as perturbations to be rejected. Convergence results are reported and a simulation example is provided to evaluate the performances of DePC.