LAUSR

Abstract To resolve the problem that greenhouse climate is difficult to model and hard to control accurately, a method of greenhouse modelling and control was developed based on Takagi-Sugeno fuzzy model (T-S model). This method is divided into two parts. In the first part, T-S model would be built based on the historical data of greenhouse instead of the physical relation. During this process, membership function was built by c-means clustering which is used to cluster the variable of historical data. The result of clustering determines the composition of both antecedent and consequence of the T-S fuzzy rules, and the parameters of T-S rules would be solved by recursive weighted least squares algorithm. In the second part, model predictive control is adopted with 3 step predictive length. According to simulations, the proposed modelling and control method give good performance and the result close to the set value.