The application of multichannel signal processing algorithms in front-ends of hands-free speech communication systems facilitates a better extraction of a desired source signal and allows for a suppression of unwanted noise and interference in car interiors. In this paper, a novel superdirective beamforming design method called the Robust Least-Squares Frequency-Invariant Beamformer (RLSFIB) is investigated for microphone arrays in cars. The RLSFIB design offers superior spatial selectivity at low frequencies even for small array apertures and some variability in microphone positioning. To cope with the high sensitivity inherent to superdirectional beamformers, this method directly incorporates a white noise gain (WNG) constraint for controlling the beamformer sensitivity. The resulting problem is formulated as a convex optimization problem which can be solved by conventional iterative solvers and the convexity guarantees a globally optimal solution. As the WNG constraint can be chosen freely, the method can match desired robustness levels, and therefore can be adapted to given prior knowledge on microphone mismatch, positioning errors, and microphone self-noise. Design examples for typical car requirements will be presented.
Robust superdirectional beamforming for hands-free speech capture in cars
2009
4 Seiten, 6 Bilder, 9 Quellen
Aufsatz (Konferenz)
Englisch
Sector-based detection for hands-free speech enhancement in cars
Tema Archiv | 2006
|CARS ultrafast light source is hands-free
British Library Online Contents | 2009
|