Nonlinear constrained predictive control applied to a coupled-tanks apparatus

The focus of the paper is the development and application to experimental e quipment of fast constrained predictive control algorithms based on feedbac k linearisation. Rather than use quadratic programming (QP) optimisation is performed, deploying a computationally much cheaper alternative based on i nterpolation and linear programming (LP). Despite its undemanding computati onal nature, this algorithm is found to perform well both in simulation, an d when applied to an actual couple-tanks rig. The advantages of the algorit hm are further illustrated by comparison with PID control.