Implementasi hidden markov model untuk aplikasi pengenalan ucapan sebagai kendali gerak robot mobil
Abstract
Robot is an electronic or mechanical equipment capable of performing a series of actions automatically. In the world of robotics, control systems is a very important part that serves to control the movement or navigation in a robot. Sound based mobile control system is an alternative voice control that is quite efficient. However, voice recognition process is not easily done by a machine. It needs a machine learning method that can be used to perform the voice feature extraction by studying the characteristics of a previous vote. This study uses Mel-Frequency Cepstral Coefficient (MFCC) to extract speech signal and HMM (Hidden Markov Model) for modeling the speech signal. Result of the experiment from the whole system performance speech recognition is 96,4% for people that have been inputted in the database, and 91,2% for people which have not been inputted in database. The mobile robot system test results showed that the system has worked well and has an average response time to move after the speech recognition process around 0,46 seconds.
Collections
- UT - Computer Science [2236]