Voice Response Translator (VRT): Support for Prototype Development and Results of Initial Field Testing

The VRT, which uses voice-recognition technologies developed in the former Soviet Union, requires users to "program" or "train" the unit to their voice patterns for specific "trigger" phases. When spoken, the VRT will respond to the trigger phrase with an audio (recorded human voice) translation in a complete command or sentence in the selected language. Because the device uses voice recognition, success in the field is dependent on the user saying the trigger phrases with the same inflection and volume as recorded at the time of programming. Over the 4-year period of development and evaluation, the number of users per device has increased to 8, and the number of phrases has expanded to approximately 200. The situations/events for which relevant translations are provided include initial greetings, crowd control, field interviews, victim interviews, medical assistance, domestic issues, lost children, traffic stops, and driving under the influence. This report provides details on software/hardware improvements, as well as the introduction of the device to several military units and the Coast Guard. One recommendation from the field evaluations is that consideration be given to developing an instructional video that demonstrates how to hold, use, speak into, and control the device. This video would be particularly useful for potential users who may not have the benefit of one-on-one instruction or who might need refresher training. Recommendations are offered for future work. 4 tables, 5 figures, and appended details on the use and evaluation of the VRT