Multilingual speech recognition extended to cars
In-car entertainment and communication systems are becoming so sophisticated and complex that they can distract motorists from their main job - driving safely. A project funded by the European Union as part of SpeechDat is working on a hands-free, speech-operated device that (in the absence of a willing partner) will allow drivers to tell their in-car devices what to do without taking their concentration off the road. Seventeen European countries have been collaborating on the Telematics Applications project SpeechDat-Car (Speech database for voice-driven teleservices and control in automotive environments) since April 1998. It is a huge project involving the collation of millions of voice samples to produce language databases for nine European languages, and is expected to finish in October 2000. According to the Commission project officer responsible, 'If you have to collect languages for so many countries, you need to involve partners from each of the Member States involved. This requires large consortia which are difficult to manage. But if they are managed well they produce networks and links between people involved, which is one of the wider aims of the European Commission.' To bring these databases to life, another project, VODIS (Advanced speech technologies for voice-operated driver information systems) has been developing an easy to use voice-based interface for telematics services in cars - a prototype of which was successfully demonstrated at the 1998 Telematics Conference in Barcelona. The VODIS II project will broaden the scope of the VODIS programme by introducing two new languages - Italian and English.