An SDP-based TDOA Localization Method for Wireless Sensor Networks

The accuracy of localization is exceedingly important since high accuracy positioning information can benefit location-based services. Non-line-of-sight (NLOS) errors are one of the most important factors affecting the accuracy of localization. To improve the positioning accuracy, it is necessary to reduce the impact of the NLOS errors. This letter proposes a novel semidefinite programming (SDP) optimization method to estimate source node position based on time-difference-of-arrival (TDOA) measurements which can mitigate NLOS errors. The main contributions of this letter include three points. First, this method does not require distinguishing NLOS connections from LOS connections, which is often difficult to be realized perfectly in practice. By formulating the original TDOA-based source localization problem as an SDP problem, the nonlinear nonconvex problem is solved. Second, new constraints about geometry and time are introduced to enhance localiztion accuracy. Third, The simulation results show that the proposed method outperforms several existing schemes in both mild and severe NLOS conditions.