ROBUST SPARSITY-BASED DEVICE-FREE PASSIVE LOCALIZATION IN WIRELESS NETWORKS

Abstract

As an emerging technique with a promising application prospect, the device-free passive localization (DFPL) technique has drawn considerable research efiorts due to its ability of realizing wireless localization without the need of carrying any device and participating actively in the localization process. Recent technological achievements of the DFPL technique have made it feasible to realize location estimation using the received signal strength (RSS) information of wireless links. However, one major disadvantage of the RSS-based DFPL technique is that the RSS measurement is too sensitive to noise and environmental variations, which incur the misjudgment of shadowed links and degradation of localization performance. Based on the natural sparsity of location flnding in the spatial domain, this paper proposes an environmental-adaptive sparsity-based localization method for the DFPL problem in the existence of model mismatch. The novel feature of this method is to adjust both the overcomplete basis (a.k.a. dictionary) and the sparse solution using a dictionary learning (DL) technique based on the quadratic programming approach so that the location solution can better match the changes of the RSS measurements between the node pairs to the spatial location of the target. Moreover, we propose a modifled re-weighting l1 norm minimization algorithm to improve reconstruction performance for sparse signals. The efiectiveness of the proposed scheme is demonstrated by experimental results where the proposed algorithm yields substantial improvement for localization performance.