Robust fuzzy model predictive control of an overhead crane

The method of controlling an overhead crane with respect to the variation of operating conditions and control constraints is developed using a model predictive control (MPC) and fuzzy interpolation applied in linear parameter varying (LPV) approach to crane dynamic modelling. The proposed control approach is based on the assumption that operating conditions vary within the known range of scheduling variables, and the parameters of a crane dynamic model can be interpolated by a quasi-linear fuzzy model designed through utilizing the P1-TS fuzzy theory. Hence, a crane dynamic is approximated through interpolation between a set of local linear models determined through identification experiments at the local operating points selected within the bounded intervals of scheduling variables. For the modelling assumptions, the control algorithm is developed based on a generalized predictive control (GPC) procedure taking into consideration the constraints on sway angle of a payload and control signal. Feasibility and applicability of the proposed control technique were confirmed during experiments carried out on a laboratory-scaled overhead crane. The results of experiments are presented and compared with performances of a fuzzy logic-based scheduling control scheme.