A nonlinear distributed model predictive scheme for systems based on Hammerstein model

Abstract

A distributed model predictive control (DMPC) scheme for systems based on the Hammerstein structure is proposed in this work. To deal with the nonlinearity of the Hammerstein model, the DMPC problem is reduced to an optimization problem on the intermediate variable and, subsequently, the inverse of the nonlinear block is considered to find the corresponding control inputs. Also, a sub-optimality bound of the method is presented. To illustrate our approach, we consider a nonlinear system controlled in a distributed manner by two agents that exchange information and make cooperative decisions. The efficiency of the proposed DMPC is demonstrated through a simulation example where it is compared to the corresponding centralized approach.