A new sparse reconstruction algorithm for device-free localization with sensor network

Device-free localization (DFL) is an emerging technology for localizing targets by monitoring the changes in the radio frequency (RF) attenuation field of an area where a wireless sensor network is deployed. Notably, this technology does not require the targets to participate in the localization effort by carrying any electronic device. Considering that the targets typically distribute in the interesting area sparsely, this paper presents an algorithm named advanced expectation conditional maximization either (ECME) thresholding pursuits (AEMTP) to realize DFL based on the compressed sensing (CS) theory. The proposed AEMTP algorithm utilizes the received signal strength (RSS) measurements of wireless links in the sensor network to reconstruct a shadow fading image and locates the targets on the none-zero pixels. The AEMTP algorithm introduces a greedy strategy that each iteration detects a support set F on the base of the ECME iteration. Then the sparse image is reconstructed by solving a truncated least-squares problem on the support set F. The experimental results reveal that the localization accuracy of the AEMTP algorithm could reach to around 0.51m by using 20 nodes in a 5m×5m square area.