Auto-generating fuzzy system modelling of physical systems
Osama Hassanein, Sreenatha Gopalarao Anavatti, Hyungbo Shim, Shaaban Ali Salman · 2015
Nonlinear system identification has gained importance over the years as enhancement tool that can improve control design and performance significantly. This is particularly true for systems with non-linearity and unmodelled disturbances. This paper proposes an Auto-Generating Fuzzy System Modelling mechanism (AGFSM) with online tuning capability. The proposed mechanism offers a universal black-box modelling tool for any system, linear or nonlinear, regardless of any prior knowledge of the physical relationship inside the system or the system behaviour. The proposed mechanism comprises of two phases, a structure-generating phase and a parameter-learning phase. Structure generating phase is based on the entropy measure used to control the model accuracy. Parameter learning phase is based on supervised learning algorithms using the back propagation algorithm. The proposed AGFSM mechanism is used to develop the models of both linear and nonlinear systems using input-output data.