Voice-based Human-Machine Interaction Modeling for Automated Information Services

R Maskeliunas,V Rudzionis,K Ratkevicius

doi:10.5755/j01.eee.110.4.300

Abstract

Voice based human-machine dialogs are becoming more and more important part of informative services. The implementation of voice dialogs enables to realize some of the aims of telecommunication services more successfully and efficiently. The main aim is to enable the communication according the principle "anytime-anywhere". The importance of voice dialogs is also caused by the fact that principle "anytime-anywhere" often could be realized only using mobile and portable devices. Those devices typically have small keyboards and screens and hence voice based interface has advantages over traditional keyboard and screen based interface. The paper presents the model of multimodal interface which core element is the recognition of voice commands. The model targets the informative services provided by the Lithuanian medical and social security enterprises. Paper shows that recognition accuracy of Lithuanian voice commands could be increased significantly if the foreign language which has closer to Lithuanian phonetic structure engine is adapted. Ill. 2, bibl. 11, tabl. 1 (in English; abstracts in English and Lithuanian).http://dx.doi.org/10.5755/j01.eee.110.4.300

Highlights

The main aim of telecommunications is to bring people thousands miles apart, anytime, anywhere together to communicate as if they were having a face-to-face conversation in a ubiquitous tele-presence way
In our previous experiments we showed that proper selection of phonetic transcriptions enables to achieve high enough recognition accuracy of Lithuanian voice commands using foreign language speech engine
We showed that selecting proper optimization procedure for the selection of phonetic transcriptions may lead to the significant improvement of the recognition accuracy

Summary

Introduction

The main aim of telecommunications is to bring people thousands miles apart, anytime, anywhere together to communicate as if they were having a face-to-face conversation in a ubiquitous tele-presence way. One key component necessary to reach this main aim is the technology enabling usual communication by voice This means the use of automatic speech recognition [1]. An IVR (Interactive Voice Response) based systems can be used to automate a wide range of services and data requests. When implementing IVR systems using mobile devices the advantages of speech recognition based interfaces becomes even more evident. Very important characteristic of voice based interfaces is the dependability of the phonetic, syntactic and lexical properties of the language spoken by the user. This means that it is impossible to move technologies developed for the recognition of one language for the recognition of another automatically. In our previous studies the advantages of such method and its possible uses were established [3, 4]

Voice based HMI for automated information services

Experimental evaluation of speech recognition accuracy

Du Trys Keturi Penki Šeši Septyni Aštuoni Devyni

String size

Findings

Conclusions