Open AccessSPEECH RECOGNITION USING SOM AND ACTUATION VIA NETWORK IN MATLAB
Author(s) -
A. Subash Chandar,
S. Suriyanarayanan,
M. Sabarimalai Manikandan
Publication year - 2014
Publication title -
international journal of electronic signal and systems
Language(s) - English
Resource type - Journals
ISSN - 2231-5969
DOI - 10.47893/ijess.2014.1166
Subject(s) - computer science , speech recognition , matlab , datagram , fast fourier transform , microcontroller , noise (video) , speaker recognition , protocol (science) , computer hardware , artificial intelligence , computer network , network packet , operating system , algorithm , image (mathematics) , medicine , alternative medicine , pathology
This paper proposes a method of Speech recognition using Self Organizing Maps (SOM) and actuation through network in Matlab. The different words spoken by the user at client end are captured and filtered using Least Mean Square (LMS) algorithm to remove the acoustic noise. FFT is taken for the filtered voice signal. The voice spectrum is recognized using trained SOM and appropriate label is sent to server PC. The client and the server communication are established using User Datagram Protocol (UDP). Microcontroller (AT89S52) is used to control the speed of the actuator depending upon the input it receives from the client. Real-time working of the prototype system has been verified with successful speech recognition, transmission, reception and actuation via network.