engleski

Decoupling Multivariable GPC With Reference Observation

This paper deals with an algorithm for decoupling multivariable systems based on generalized predictive control. Proposed control algorithm uses adaptive error weighting in cost function and reference change observation. From filtered reference it auto-tunes weight coefficients in order to reduce decoupling between control loops. The algorithm is applied to greenhouse climate control ; obtained simulation results show its efficiency.

Predictive control; MIMO GPC; decoupling

nije evidentirano