No abstract

With the spread of mobile communication networks, location dependent services are gaining popularity all over the world. Obviously, as its name implies, a location dependent service needs to know (more practically speaking, estimate) the location of the mobile station (MS) before it can offer any service at all. In this paper, we address the issue of mobile positioning (i.e. estimating the location of an MS) given the measured distances between the MS and its nearby base stations (BS's). The measurements, which are in general corrupted with measurement noise and NLOS (non-line-of-sight) error, can come from the time of arrival (TOA), the time sum of arrival (TSOA), or the time difference of arrival (TDOA). The NLOS error is the major cause that degrades the accuracy of mobile positioning. Assuming the knowledge of its statistic model, however, we propose a scheme that greatly reduce its effect. The results on MS location estimation thus obtained are comparable to the Cramer-Rao lower bounds (CRLB's). Regardless of the measurement types (i.e. TOA, TSOA, or TDOA), the proposed scheme computes the MS location in a unified way.