Efficient implementation of simplified model predictive control

Abstract

Listen

Translate article

One of the recent digital control algorithms is known as the simplified model predictive control (SMPC) algorithm. The SMPC algorithm has many attractive features, such as zero offset, a single tunable parameter, some dead time compensation, and others. One disadvantage of SMPC compared with PID (proportional-integral-derivative) control is the computational effort; it involves N + 2 multiplications (and relatively inexpensive N additions) where N is the number of sampling periods in the settling time. This paper describes a simple yet effective implementation of SMPC that significantly reduces the required computational effort. Furthermore, the proposed modification improves the performance of SMPC for many applications of practical importance.