Understanding Model Predictive Control, Part 4: Adaptive, Gain-Scheduled and Nonlinear MPC

This video explains the type of MPC controller you can use based on your plant model, constraints, and cost function. - Model Predictive Control Toolbox: http://bit.ly/2xgwWvN- - What Is Model Predictive Control Toolbox?: http://bit.ly/2xfEe2M The available options include the linear time-invariant, adaptive, gain-scheduled, and nonlinear MPC. - Adaptive MPC Design: http://bit.ly/2Gv4pcX - Gain-Scheduled MPC Design: http://bit.ly/2GwnriT If you have a linear plant model and your MPC problem has linear constraints and a linear cost function, then you can use linear time-invariant MPC to control your system. - How to Design an MPC Controller with Simulink and Model Predictive Control Toolbox: http://bit.ly/2Gvv0qe - Adaptive MPC Design with Simulink and Model Predictive Control Toolbox: http://bit.ly/2GsL5Nu An optimization problem with these properties is a convex one, and you can use many types of there are numerical methods and software to solve it. If your system is nonlinear, but it can be approximated by linear models at operating points of interest, then you can use adaptive or gain-scheduled MPC. In adaptive MPC, a linear model is computed on the fly as the operating conditions change. In gain-scheduled MPC, the linearization is performed offline at the operating points of interest. If your plant is highly nonlinear, these options probably won’t provide satisfactory performance. In that case, you can use nonlinear MPC. In addition, you can use nonlinear if you have a linear plant model but either the constraints, the cost function, or both are nonlinear.