Blind Blind Sampling Rate Offset Estimation for Wireless Acoustic Sensor Networks through Coherence Drift Estimation
Proposal for a Master Thesis
Topic:
Blind Blind Sampling Rate Offset Estimation for Wireless Acoustic Sensor Networks through Coherence Drift Estimation
Description:
In Wireless Acoustic Sensor Networks (WASNs) spatially distributed nodes allow for multi-channel array signal processing tasks, e.g., speech enhancement, Blind Source Separation (BSS), Localization, Acoustic Echo Cancellation (AEC) etc. As opposed to single microphone arrays that capture a sound field only locally, WASNs provide different perspectives on the acoustic scene. An important area of application of WASNs are smart home environments which should assist residents in their daily life.
However, the individual nodes are not always sampled synchronously due to clock imperfections. Thus, already a minor offset or mismatch in sampling rate causes a linearly increasing time delay corresponding to a phase shift in the STFT-domain between the microphone signals of different nodes, which may severely degrade the performance of signal processing algorithms. Hence, the Sampling Rate Offset (SRO) must be estimated and compensated for.
Professor:
Prof. Dr.-Ing. Walter Kellermann
Supervisior:
Matthias Kreuzer, M.Sc., room 05.018 (Cauerstr. 7), matthias.kreuzer@fau.de
Available:
Immediately