A new cooperative received signal strength-based localization algorithm is proposed which employs relative error estimation and semidefinite programming (SDP). First, the log-normal shadowing RSS measurement model is transformed into an equivalent multiplicative model. Then, a relative error estimation criterion is used with this model to develop a nonconvex estimator to approximate the maximum likelihood solution. Finally, semidefinite relaxation is applied to the nonconvex estimator to obtain an SDP estimator. The proposed algorithm is first derived for noncooperative RSS-based localization and then extended to the cooperative case. The Cramer-Rao lower bound is derived for cooperative RSS-based localization. Performance results are presented, which demonstrate that the proposed SDP estimator provides a significant improvement over existing localization methods.

Location information of sensor nodes is of vital importance for most applications in wireless sensor networks (WSNs). This paper proposes a new range-free localisation algorithm using support vector machine (SVM) and polar coordinate system (PCS), LSVM-PCS. In LSVM-PCS, two sets of classes are first constructed based on sensor nodes’ polar coordinates. Using the boundaries of the defined classes, the operation region of WSN field is partitioned into a finite number of polar grids. Each sensor node can be localised into one of the polar grids by executing two localisation algorithms that are developed on the basis of SVM classification. The centre of the resident polar grid is then estimated as the location of the sensor node. In addition, a two-hop mass-spring optimisation (THMSO) is also proposed to further improve the localisation accuracy of LSVM-PCS. In THMSO, both neighbourhood information and non-neighbourhood information are used to refine the sensor node location. The results obtained verify that the proposed algorithm provides a significant improvement over existing localisation methods.

To improve localization accuracy and reduce system costs, a new range-free localization algorithm for wireless sensor networks using connectivity and received signal strength (RSS) rank is proposed. An unknown node is first localized in an initial residence area according to the connectivity constraints. Then the residence area is refined using the RSS rank vector. The RSS rank vector of an unknown node is formed from the ranks of the RSS values obtained from neighboring anchors. The estimated location of the unknown node is the centroid of the refined residence area. The proposed localization algorithm uses a grid scan approach to avoid complex geometric computations. This algorithm is improved by using an adaptive strategy to determine the grid size. The performance of these methods is examined via analysis and simulation. The results obtained verify that the proposed approaches provide better localization accuracy than competing algorithms in the literature.

Passive source localization is an important issue as there are numerous applications including search and rescue and public safety. Hybrid passive source localization combining time difference of arrival (TDOA) and received signal strength difference (RSSD) is investigated in this paper. First, a TDOA and RSSD based linear weighted least-squares (WLS) estimator for passive source localization is presented. Second, the maximum likelihood (ML) estimator and Cramer-Rao lower bound (CRLB) for the hybrid localization scheme are derived as performance benchmarks. Numerical results are presented which demonstrate that the proposed hybrid localization method provides significantly better accuracy than using only the TDOA.