Please use this identifier to cite or link to this item:
|Title:||Compressive beamforming accelerated with the kronecker array transform|
|Abstract:||The problem of acoustic scene description with sensor ar- rays is to determine the number and location of (usually few) sound sources present in a (possibly noisy) sound scene from measurements of the wave field with a mi- crophone array. Conventional beamforming is the most usual method to extract the sources’ direction-of-arrival and emitted signal, even though it is characterized by low spatial resolution. The compressive beamforming (CB) method asserts that spatially sparse signals can be recovered from arrays with reduced number of sensors by solving a convex minimizati- on problem. However, despite the fact that the compressi- ve sensing framework applied in CB offers computational efficiency compared to other sparsity promoting methods, its iterative algorithm is still very time consuming when compared with conventional beamforming. In the quest for a real-time implementation of CB, we present the Kronecker Array Transform (KAT) to speed up the bott- leneck of the CB algorithm, namely, the matrix-vector product calculation, which requires as trade-off for con- siderable calculation speed up the use of a sensor array with separable geometry.|
|Appears in Collections:||FEEC - Artigos e Outros Documentos|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.