Development of a new type of recurrent co-active neuro-fuzzy system for system identification

In this paper, the development of a new type of recurrent co-active neuro-fuzzy system is presented, together with its application to system identification. Hybrid learning based on fuzzy clustering algorithm and the steepest-descent method is used to train the proposed neuro-fuzzy system. The experimental case studies refer to the identification of a simulated plant and the evaporator system of the evaporation station from a sugar factory. The identification is performed using the proposed recurrent co-active neuro-fuzzy system. The obtained results demonstrate the efficiency of the approach.