THE HARTES CARLAB: HARDWARE IMPLEMENTATION AND ALGORITHM DEVELOPMENT
Résumé
In the last decade Car Infotainment Systems (CIS) have been gaining great attention by the scientific and industrial community: in this context, within the European hArtes Project, an advanced CIS (ACIS) has been designed. The system offers several functionalities employing professional audio equipment and PCs able to manage different I/O audio channels and to provide a large computing capability for complex audio algorithms. The overall architecture is based on NU-Tech platform which manages the whole system from professional equipment to audio streaming and processing. The system has been therefore devised as a real audio laboratory (hArtes CarLab) for audio algorithm exploration and validation, providing a remote access to all the system functionalities. In this paper, starting from the hardware description, a complete set of algorithms to enhance audio reproduction, hands-free communication,and interactivity through speaker and speech recognition features is discussed in relation to the NU-Tech framework. INTRODUCTION In the last decade Car infotainment (i.e. the combination of information with entertainment features) systems have attracted many efforts by industrial research because car market is sensible to the introduction of innovative services for drivers and passengers. The need for an Advanced Car Infotainment System (ACIS) has been recently emerging, able to handle issues left open by CIS systems already on the market, and to overcome their limitations. Due to traffic congestion and growing distance from home to workplace, people spend more and more time in car, that hence becomes an appealing place to do many common activities such as listening to music and news, phone calling and doing many typical office tasks. Narrowing our focus to audio, the key role of the CIS is that it lets the driver concentrate on the road and at the same time it manages many different processing functions such as high quality music playback, hands-free communication, voice commands, speaker recognition, etc. Moreover, from the signal processing point of view, the wideband nature of the audio signal adds complexity while the requested quality calls for high precision signal processing,