Implementation of Speaker Independent Speech Recognizer for Embedded system

In this paper, the speaker independent speech recognition is used to control embedded system. The factor of speech feature vector is based on LPCC which is modeling a vocal tract and selected 20 words for system control are derived from sampling of speech recognition feature and Data Base for comparison of pattern in HMM through training of Baum-Welch Algorithm. We used a microprocessor, 80c196kc for the efficient conduct of speech recognition processing. We made an each recognition experiment both of indoors outdoors to find out the difference of recognition rate by noise environment and resulted in each 95.42% and 93 % recognition.