Development of an Advanced Printed Circuit Board Technology Using Liquid Crystal Polymers

Eliminating noise from communication channels is important for both military and commercial applications. Conventional approaches to noise elimination use only one sensor, i.e. the ordinary microphone. Although noise can be eliminated, the distortion ofthe speech signal due to signal processing actually reduces speech recognition rates even compared to the original noisy speech signal.In this Phase II proposal, Intelligent Automation, Inc. (IAI) will develop hardware that fuses information from different forms of speech sensors. In particular, we will fuse information from a new type of microphone called a Glottal ElectromagneticMicropower Sensor (GEMS) microphone developed by LLNL, camera images quantifying lip motion, bone conduction microphones, and ordinary microphones. The work will require refinement of our Phase 1 camera sensor for quantifying lip motion, and refinement ofthe GEMS sensor developed by LLNL, as well as further development of our signal fusion methodology. Our Phase I results have clearly demonstrated the feasibility of our proposed approach, however substantial additional algorithmic and hardware developmentwork is still needed.