Leveraging Compressive Sensing for Multiple Target Localization and Power Estimation in Wireless Sensor Networks

Peng QIAN  Yan GUO  Ning LI  Baoming SUN  

IEICE TRANSACTIONS on Communications   Vol.E100-B   No.8   pp.1428-1435
Publication Date: 2017/08/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.2016EBP3333
Type of Manuscript: PAPER
Category: Network
multiple target localization,  transmitting power estimation,  sensing matrix reconstruction,  compressive sensing,  wireless sensor networks,  

Full Text: PDF(847.4KB)>>
Buy this Article

The compressive sensing (CS) theory has been recognized as a promising technique to achieve the target localization in wireless sensor networks. However, most of the existing works require the prior knowledge of transmitting powers of targets, which is not conformed to the case that the information of targets is completely unknown. To address such a problem, in this paper, we propose a novel CS-based approach for multiple target localization and power estimation. It is achieved by formulating the locations and transmitting powers of targets as a sparse vector in the discrete spatial domain and the received signal strengths (RSSs) of targets are taken to recover the sparse vector. The key point of CS-based localization is the sensing matrix, which is constructed by collecting RSSs from RF emitters in our approach, avoiding the disadvantage of using the radio propagation model. Moreover, since the collection of RSSs to construct the sensing matrix is tedious and time-consuming, we propose a CS-based method for reconstructing the sensing matrix from only a small number of RSS measurements. It is achieved by exploiting the CS theory and designing an difference matrix to reveal the sparsity of the sensing matrix. Finally, simulation results demonstrate the effectiveness and robustness of our localization and power estimation approach.