In-car speech and audio processing - some experiments within hArtes project

This paper describes part of the work dealing with in-car speech and audio applications achieved within the European FP6-IST hArtes project. After a brief introduction describing the hArtes project framework together with the targeted scenarios, the in-car hardware platform will be presented, followed by an overview of the different algorithmic solutions. The applications are covering speech and audio processing for enhanced communications through hands-free telephony, and preprocessing for advanced speech processing modules such as speech recognition and speaker authentication. Then a description of the hArtes integration process applied to the proposed algorithmic solutions will be given, followed by some experimental results. Finally some perspectives will be presented for the next steps towards the integration on a dedicated hardware.