Cooperative distributed model predictive control for nonlinear systems

In this paper, a distributed controller is presented that can stabilize nonlinear systems. A novel nonlinear nonconvex optimizer is proposed that improves the objective function and is feasible at every iterate. The optimization uses gradient projection and converges to stationary points. The proposed optimization does not require a coordination layer, and hence the controller is truly distributed. Asymptotic stability is established for the controlled system, and an illustrative example is presented.