Fuzzy membership function optimization for system identification using an extended Kalman filter

The generation of membership functions for fuzzy systems is a challenging problem. In this paper, we use an extended Kalman filter to optimize the membership functions for system modeling, or system identification. We describe the algorithm and then show the result as sub-optimal novel method of system identification. The ideas described in this paper are illustrated for system identification of a nonlinear dynamic system of a permanent magnet synchronous motor. The other interesting observation made is that the proposed system acts as a noise-reducing filter. We demonstrate that the extended Kalman filter can be an effective tool for identifying the parameters of a fuzzy system model.