Mehmet Ali Çavuşlu, Cihan Karakuzu
6th International Advanced Technologies Symposium (IATS’11) Proceedings CD, vol 3, p.115-119, Elazığ, Türkiye, 16-18 Mayıs 2011
Abstract: In this study, wavelet fuzzy system (WFS) together with its PSO-based learning is hardware implemented on FPGA. Floating point number format is used for the implementation considering its precision and dynamics. Although floating point numbers consume more hardware source than the other number format, WFS implementation has been achieved with only 10% hardware sources. A mathematical approximation for implementing of wavelet membership functions has been proposed and exploited. Proposed hardware implementation method has been experimentally inspected on a benchmark system identification problem. Obtained results show that implemented WFS has carried out good performance not only training data but also tested data on the system identification problem.