The acquisition and localization of desired sounds is essential in a broad range of acoustical applications, such as surveillance, security monitoring, heath-care monitoring, video-conferencing systems. The desired sounds that are received by an acoustic sensor are degraded by room reverberation, background noise, and other interferences. This degradation may lead to total unintelligibility of the desired sounds (speech and audio) and decreases the performance of automatic speech recognition systems. Therefore, Acoustic Signal Processing (ASP) techniques that capture desired sounds or their source locations from low-quality signals are of major importance. While state-of-the-art ASP algorithms are able to reduce noise, the development of practical algorithms that can eliminate the degradations caused by reverberation in a noisy environment have not yet emerged. Furthermore, it should be noted that many, if not all, existing ASP techniques fail completely or experience a dramatically reduced performance when noise and reverberation are present. The main objective of this project is to develop advanced ASP algorithms that will allow the capture of desired sounds and their source locations in strongly adverse environments. We anticipate that this objective can be achieved by using distributed acoustic sensors and sensor arrays that measure the sound pressure, as well as the particle velocity. In addition, we propose to exploit statistical room acoustic theory in the development of advanced ASP algorithms.
Field of science
- /humanities/languages and literature/linguistics/phonetics
- /engineering and technology/electrical engineering, electronic engineering, information engineering/electronic engineering/signal processing
Call for proposal
See other projects for this call