For each node comprising many channels in the wireless acoustic sensor networks (WASNs), the signals from different channels both have the spatial-temporal correlation property. The two-dimensional discrete cosine transform (2D-DCT) is introduced to decorrelate and compute the sparse representation of the signals. Then, only the first K largest 2D-DCT coefficients and the binary sequences consisting of 0 and 1 to indicate the indices of these coefficients are transmitted. The proposed method reduces the number of channels transmitted per node and yet the raw signals can still be recovered accurately at the fusion center. The proposed method is independent of the specific beamformer compared to other distributed algorithms to reduce the communication overhead, and the experimental results demonstrate its validity.
(See document for exact affiliation information.)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
