Robust constrained model predictive control for nonlinear systems: a comparative study

We compare two approaches to controlling nonlinear systems using model predictive control (MPC) techniques. In the first approach, we use an inner feedback loop to linearize the nonlinear plant and use the resulting linear model to synthesize a controller strategy based on standard MPC techniques. The nonlinear constraints resulting from the feedback linearization are handled using an iterative method. In the second approach, we approximate the nonlinear system by a linear time-varying system and design a stabilizing receding horizon state-feedback control law using optimization techniques based on linear matrix inequalities.