ENSEMBLE MLP NETWORKS FOR VOICES COMMAND CLASSIFICATION TO CONTROL MODEL CAR VIA PIFACE INTERFACE OF RASPBERRY PI

Authors

  • Narissara Eiamkanitchat
  • Nontapat Kuntekul
  • Phasit Panyaphruek

Keywords:

Thai voices command, MLP neural network, Speech recognition, Classification

Abstract

This research, exploration displays the aftereffects of utilizing the blend of the multi-layer
perceptron network system to classify Thai speech. The parameters of the training process are used in the
mobile application to using Thai voice commands to control the model car. The PiFace interface of the
Raspberry Pi is attached to the model car for receiving the command from mobile and control the model car.
The 1,000 Thai voice commands of both men and ladies are used as the training set in the experiment. The
preliminary experiments have been done to find the best possible structure of the classification model, and
the appropriate proportion of classes in the training set. From the experiment results using 1 network for one
voice command, the average accuracy of the classification results in the environment without noise is higher
than 80%, which considered favorable in the speech recognition field of study.

Downloads

Published

2016-11-30

How to Cite

Narissara Eiamkanitchat, Nontapat Kuntekul, & Phasit Panyaphruek. (2016). ENSEMBLE MLP NETWORKS FOR VOICES COMMAND CLASSIFICATION TO CONTROL MODEL CAR VIA PIFACE INTERFACE OF RASPBERRY PI. GEOMATE Journal, 13(37), 9–15. Retrieved from https://geomatejournal.com/geomate/article/view/1498

Most read articles by the same author(s)