VODIS

Project Summary

The main objective of the project is to integrate and further develop the enabling technologies for the design and implementation of voice-operated human-machine-interface (HMI) for applications inside the automobile. The goal of such a vocal interface is to enhance the
usability of newly developed driver assistance-and-information systems in the sense that it facilitates the access to the information provided, due to the fact that spoken language is the most natural form of human interaction.

In order to address actual needs, the work will be focused on the voice-operation of a commercially available state-of-the-art driver information system, namely the BERLIN RCM303A from Bosch, that presently integrates a route guidance (navigation) computer with traffic-messaging (RDS/TMC), mobile telephone (GSM) and more conventional equipment as
car-radio and audio/video entertainment sources, thus demanding a high amount of driver interaction, both for the command-and-control functions and database queries. The speech technologies and evaluation methods developed within the project are of generic nature and can be used for a multiplicity of systems and services beyond the particular application,
not only in the automobile but also in other acoustically hostile scenarios.

The main core of the project is a speaker-independent noise robust phoneme recognizer with enhanced spotting-and-rejection capabilities and some level of language knowledge in form of finite state grammars. Given the noisy acoustical environment inherent to the driving scenario and the noise-sensitiveness of phonetically-based recognizers, considerable
effort will be devoted to the speech signal enhancement. This workpackage will mainly be managed by the Aachen University of Technology. Bosch and Lernout & Hauspie Speech Products will provide significant contributions.

Workpackage - Speech Enhancement: 

Any hands-free application of a Voice Control Unit (VCU) comprising speech recognition and speech output has to deal with two major problems. Firstly, due to the coupling between loudspeaker and microphone the loudspeaker signal is echoed back to the microphone. Secondly, due to the relatively large distance from the speakers mouth to the microphone the microphone signal might be severely degraded by ambient noise. Therefore, countermeasures, i.e.methods of Speech Enhancement, are necessary with respect to error rates of the speech recognizer. The workpackage Speech Enhancement  combines echo
cancellation and noise reduction tasks in a real time DSP based implementation. Thus, the workpackage comprises the following tasks:

• Acoustic Echo Control
• Noise Reduction
• Combination of Echo Control and Noise Reduction
• Combined Optimization of Enhancement and Robust Recognition
• Development of a Real Time System