A MIMO fuzzy model-based controller

Fuzzy model-based controllers offer a means of controlling non-linear MIMO systems. The main problem with implementing these systems is that the normal model formulations become increasingly difficult to invert as the number of controlled outputs increase. A new fuzzy model formulation is presented in this paper which avoids this problem. Details are also given of a new identification scheme which allows a priori information about the system to be taken into account. A new MIMO controller algorithm is presented which calculates the required control actions by analytically inverting a MIMO fuzzy model at each sample interval. Finally, some simulation results showing the simultaneous control of pH and level using inlet base and acid flows are given as a demonstration. The simulation results show that the new model structure, identification method, and controller are capable of producing excellent results in the multi-variable control of a highly non-linear, interactive, process.