Personal Identification using Voice Recognition with Neuro Fuzzy Method
Abstract
The Experimental Power Plant (RDE) is one of the vital object that will be built by the Indonesian nation. During the operation time, RDE needs to pay attention to the nuclear security aspect. Personnel whose accessing the RDE is need to be controlled especially in the restricted area. It needs hardware systems and supporting software to ensure the nuclear security. The purpose of this research is to design personnel access system using voice recognition to support the development activity of detection and response system of RDE test facility. This is related to the development of nuclear security. Several parameters used in this research are voice samples with sampling frequency 8000 Hz and 8 bit per sample with High Pass Filter wavelet filter (HPF). We use wavelet coiflet level 2 for decomposition level and Shannon entropy is used to calculate wavelet optimization, by used that we get a characteristic vector of each speaker by value Feature vector (Shannon Entropy). We used Neuro-Fuzzy N-Input and 1 Output, with 19 class classification (19 personnel) to identify personnel. In this research, we get time about 4.7379 s for identification personnel.
- Publication:
-
Journal of Physics Conference Series
- Pub Date:
- April 2019
- DOI:
- 10.1088/1742-6596/1198/9/092005
- Bibcode:
- 2019JPhCS1198i2005S