Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - Reordering for better compressibility: Efficient spatial sampling in Wireless Sensor Networks
AU - Mahmudimanesh, M.
AU - Khelil, A.
AU - Suri, Neeraj
PY - 2010/6/7
Y1 - 2010/6/7
N2 - Compressed Sensing (CS) is a novel sampling paradigm that tries to take data-compression concepts down to the sampling layer of a sensory system. It states that discrete compressible signals are recoverable from sub-sampled data, when the data vector is acquired by a special linear transform of the original discrete signal vector. Distributed sampling problems especially in Wireless Sensor Networks (WSN) are good candidates to apply CS and increase sensing efficiency without sacrificing accuracy. In this paper, we discuss how to reorder the samples of a discrete spatial signal vector by defining an alternative permutation of the sensor nodes (SN). Accordingly, we propose a method to enhance CS in WSN through improving signal compressibility by finding a sub-optimal permutation of the SNs. Permutation doesn't involve physical relocation of the SNs. It is a reordering function computed at the sink to gain a more compressible view of the spatial signal. We show that sub-optimal reordering stably maintains a more compressible view of the signal until the state of the environment changes so that another up-to-date reordering has to be computed. Our method can increase signal reconstruction accuracy at the same spatial sampling rate, or recover the state of the operational environment with the same quality at lower spatial sampling rate. Sub-sampling takes place during the interval that our reordered version of the spatial signal remains more compressible than the original signal. © 2010 IEEE.
AB - Compressed Sensing (CS) is a novel sampling paradigm that tries to take data-compression concepts down to the sampling layer of a sensory system. It states that discrete compressible signals are recoverable from sub-sampled data, when the data vector is acquired by a special linear transform of the original discrete signal vector. Distributed sampling problems especially in Wireless Sensor Networks (WSN) are good candidates to apply CS and increase sensing efficiency without sacrificing accuracy. In this paper, we discuss how to reorder the samples of a discrete spatial signal vector by defining an alternative permutation of the sensor nodes (SN). Accordingly, we propose a method to enhance CS in WSN through improving signal compressibility by finding a sub-optimal permutation of the SNs. Permutation doesn't involve physical relocation of the SNs. It is a reordering function computed at the sink to gain a more compressible view of the spatial signal. We show that sub-optimal reordering stably maintains a more compressible view of the signal until the state of the environment changes so that another up-to-date reordering has to be computed. Our method can increase signal reconstruction accuracy at the same spatial sampling rate, or recover the state of the operational environment with the same quality at lower spatial sampling rate. Sub-sampling takes place during the interval that our reordered version of the spatial signal remains more compressible than the original signal. © 2010 IEEE.
KW - Compressed sensing
KW - Compressibility
KW - Compressive wireless sensing
KW - Permutation
KW - Reordering
KW - Spatial sampling
KW - Wireless sensing
KW - Data compression
KW - Metadata
KW - Mobile computing
KW - Optimization
KW - Sensor networks
KW - Sensor nodes
KW - Signal reconstruction
KW - Technical presentations
KW - Telecommunication equipment
KW - Ubiquitous computing
KW - Wireless sensor networks
U2 - 10.1109/SUTC.2010.30
DO - 10.1109/SUTC.2010.30
M3 - Conference contribution/Paper
SN - 9781424470877
SP - 50
EP - 57
BT - 2010 IEEE International Conference on Sensor Networks, Ubiquitous, and Trustworthy Computing
PB - IEEE
ER -