Generalized Predictive Control using Interval Type-2 Fuzzy models

The development of Interval Type-2 Fuzzy Logic Systems has brought great improvements in the non-linear system modeling domain. However, in what concerns to the development of control systems, the approaches found in literature of Type-2 Fuzzy Sets do not seem to be taking fully advantage of the advances achieved by adaptive self-tuning algorithms, already well established in both academic and industrial communities. This work presents how a controller based on Generalized Predictive Control (GPC) theory can be developed based on an Interval Type-2 Takagi-Sugeno Fuzzy Model, providing details regarding the online model training mechanisms and controller parameter's synthesis. This approach is then compared with two additional GPC implementations based on an Auto-Regressive model with eXogenous inputs (ARX) and Type-1 Takagi-Sugeno Fuzzy models. A Multiple-Input-Single-Output (MISO) Coupled Tank System will serve as benchmark system to evaluate the reference tracking capability and robustness of the controllers when subjected to different operation points and several unmodeled disturbances.