Efficient Nonlinear Model Predictive Control For Discrete System With Disturbances

Computational resources are vital especially for systems with limited computational capabilities. We propose a computationally efficient formulation for Nonlinear Model Predictive Control (NMPC) algorithm for a discrete time nonlinear system affected by additive and multiplicative disturbances. Factors which contribute to the computationally efficient formulation are: avoiding stability related terminal cost and constraints in the optimization problem formulation, varying prediction horizon online according to the cost value and use of event triggering. Simulations and analysis were carried out in nonlinear models of mechanical system and the Quadruple Tank Process (QTP). The effect of disturbance on the control algorithm was considered and solutions suggested.