Adverse-Environment Recognition of Speech


ARS - 2101

Keywords speech recognition, noise robustness, user interfaces


Start Date: 16-MAR-89 / Duration: 51 months

[ contact / participants ]


Objectives and Approach

The objective of ARS was to develop improved algorithms for medium-size vocabulary speaker-dependent speech recognition in the presence of noise, and to build a real-time demonstrator. The demonstrator was to incorporate an isolated word noise-robust recogniser, verify algorithm performance, and address the problem of speech-based person-machine dialogue as a system interface in practical applications. The application environment chosen was the car.

The requirements included a 100-word vocabulary, chosen for each language group of partners and tailored to the specific application environment. Advances were needed in terms of:

The system would be integrated in a real-time demonstrator based on a general-purpose DSP chip attached to a personal computer on a stand-alone system. Performance evaluations were first scheduled in the laboratory, using databases collected in noisy environments, to evaluate the resulting rate of correct recognition. Performance under field conditions were then to be assessed from a prototype fitted in a car and a laboratory system installed in a factory.

Progress and Results

A "baseline" system, representing the state-of-the-art at the start of the project, was defined and subdivided into modules. A multilingual database was collected in noisy environments, providing a reference point for the performance of the baseline system. Partners concentrated their efforts on developing improvements of different algorithms inside particular modules. Finally, a comparative evaluation of competing solutions resulted in the choice of a set of optimal algorithms for the final system. The various algorithms were ported onto a real-time hardware. Two demonstrators were installed inside cars for assessing their performance in real operating conditions.

Exploitation

The results of the project find applications in:


CONTACT POINT

Mr G. Babini
CSELT
Via G. Reiss Romoli 274
I - 10148 TORINO
tel: + 39/ 11-2169417
fax: + 39/ 11-2169520
telex: 220539 CSELT I

Participants

CSELT - I - C
TELECOM PARIS - F - P
MATRA COMMUNICATION - F - P
UNIVERSITY OF CAMBRIDGE - UK - P
UNIVERSIDAD POLITECNICA DE MADRID - E - P
UNIVERSITY OF KEELE - UK - P
PAGE IBERICA SA - E - P
LOGICA LTD - UK - P
LOGICA SDS LTD - UK - P
POLITECNICO DI TORINO - I - A


ST synopses home page ST acronym index ST number index
All synopses home page all acronyms index all numbers index

ARS - 2101, December 1993


please address enquiries to the ESPRIT Information Desk

html version of synopsis by Nick Cook