Neuro-fuzzy multi-model control using Sugeno inference and Kohonen tuning in parameter space

A multi-model adaptive control scheme is introduced. It makes use of Kohonen neural structures and Sugeno fuzzy inference in adapting and switching the control action of the multi-model bank respectively. The structure can be easily put in a parametrized gain scheduling framework, practical to engineering situations. A simple example of a tracking filter (radar) has been simulated with good results.