Fuzzy Filtering and its Applications
Prem Kumar Kalra, Amar Mishra, Anupam Prabhakar · IETE Journal of Research · 1998
This paper is concerned with system identification using fuzzy logic. The focus is on the blending of the classical least square error minimization approach with fuzzy logic which is used to fuzzify the model parameters. The model parameters for hourly load are used to predict the future values of the load. The results for system identification and load prediction are compared with those obtained using non-fuzzy methods. Speech data of an individual was also recorded and then used to separate the glottal wave using cepstrum analysis and homomorphic deconvolution. The same model for system identification is used to find the model parameters of the glottal system using the data for the separated glottal wave.