Demo Video Project54® Introduction CATLab Team About Project54® Supported Devices Software Cost for Project54® Installation Manual Current Users Procedure Parts Suppliers and Installers Software Distributors Theses/Dissertations Technical Reports Papers Brochures/Manuals Presentations Project54® in the Media Image Gallery Videos Overview Areas

Command-and-control SUI

The SUI uses a commercial speech recognition (SR) engine and a commercial text-to- speech (TTS) engine. The SUI gets speech input from a directional microphone on the visor which reduces the influence of sounds that are not coming from the driver of the vehicle Ð this improves SR engine performance. The SUI also uses a push-to-talk button on the steering wheel. Recognition is started when the PTT button is pressed and it is stopped when the button is released. SR performance is also improved through the introduction of grammars. Grammars describe rules that the SR engine "believes" govern utterance generation. Grammars are switched by the SUI in response to officer utterances. For example, the officer may be filling in a form in which one of the fields describes the gender of a person. The officer may say 'Gender' to signal that he/she will tell the SUI the gender of the person next. At this point the SUI loads a simple grammar which only lists the two genders and one or two other utterances (such as "cancel").

In our system human-computer interaction is command-and-control interaction: the officer utters a phrase and the SUI reacts to this utterance. The SUI may execute a command, fill in a data field or initiate data retrieval. The SUI does not initiate interaction.