LAUSR

When signals propagate along non-line-of-sight (NLOS) paths, measurement results between sensor network nodes will experience significant errors. Most of the existing time-difference-of-arrival (TDOA) localization methods with higher accuracy require prior knowledge of the NLOS environment or use complex calculation methods such as convex optimization. A prior-knowledge-free method with high efficiency is proposed. Considering that the influence of NLOS environment on TDOA localization is determined by the differences of NLOS errors between each two nodes, a balance parameter will be introduced into the measurement information of the node with the most dispersed NLOS error. This parameter is designed to minimize the variance of the array of NLOS errors, in order to achieve a compensatory effect. The node with the most dispersed NLOS error is selected by using the relationship between the cost function of approximate maximum likelihood estimation and the variance of NLOS errors. A two-step weighting method is proposed to obtain an initial estimate of the target’s coordinates for calculating the value of the balance parameter. Finally, the balance parameter is added to the TDOA measurement information to estimate the target position. Through simulations and experiments, it is proved that the proposed method has high positioning accuracy and timeliness under different NLOS conditions.