TY - GEN
T1 - Decentralized joint sparsity pattern recovery using 1-bit compressive sensing
AU - Kafle, Swatantra
AU - Kailkhura, Bhavya
AU - Wimalajeewa, Thakshila
AU - Varshney, Pramod K.
N1 - Publisher Copyright:
© 2016 IEEE.
PY - 2017/4/19
Y1 - 2017/4/19
N2 - We address the problem of decentralized joint sparsity pattern recovery based on 1-bit compressive measurements in a distributed network. We assume that the distributed nodes observe sparse signals which share the same but unknown sparsity pattern. Each node obtains measurements via random projections and further quantizes its measurement vector element-wise to 1-bit. We develop two decentralized variants of the binary iterative hard thresholding (BIHT) algorithm where each node communicates only with its one hop neighbors and exchanges its measurement information. This stage is followed by index fusion stage. For first and second algorithms, index fusion is performed at the end of and during BIHT iterations, respectively. The global estimate of the support set in both the algorithms is obtained by fusing all the final local estimates. Experimental results show that the proposed collaborative algorithms have better (or at least similar) performance compared to the centralized version.
AB - We address the problem of decentralized joint sparsity pattern recovery based on 1-bit compressive measurements in a distributed network. We assume that the distributed nodes observe sparse signals which share the same but unknown sparsity pattern. Each node obtains measurements via random projections and further quantizes its measurement vector element-wise to 1-bit. We develop two decentralized variants of the binary iterative hard thresholding (BIHT) algorithm where each node communicates only with its one hop neighbors and exchanges its measurement information. This stage is followed by index fusion stage. For first and second algorithms, index fusion is performed at the end of and during BIHT iterations, respectively. The global estimate of the support set in both the algorithms is obtained by fusing all the final local estimates. Experimental results show that the proposed collaborative algorithms have better (or at least similar) performance compared to the centralized version.
KW - 1-bit compressive sensing
KW - Binary iterative hard thresholding (BIHT)
KW - Compressive sensing
KW - Information fusion
KW - Sparsity pattern recovery
UR - http://www.scopus.com/inward/record.url?scp=85019221797&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85019221797&partnerID=8YFLogxK
U2 - 10.1109/GlobalSIP.2016.7906062
DO - 10.1109/GlobalSIP.2016.7906062
M3 - Conference contribution
AN - SCOPUS:85019221797
T3 - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings
SP - 1354
EP - 1358
BT - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2016 IEEE Global Conference on Signal and Information Processing, GlobalSIP 2016
Y2 - 7 December 2016 through 9 December 2016
ER -