Performance improvement of TOA localization using IMR-based NLOS detection in sensor networks

The sensor network with many small terminals communicating each other can cooperatively collect wider and variety of information and is expected to provide advanced services. To serve such new applications, the position information of targets is essential and the exact localization techniques are necessary. Time of arrival (TOA) localization is one of the popular schemes because of its high estimation accuracy. In TOA scheme, the target node of unknown position sends a signal to the anchor nodes of known position, from the arrival time stamps of the received signal the distance between the target and the anchor nodes is derived, and then, the estimated position is obtained. However, a non-line-of-sight (NLOS) environment between the target and the anchor nodes causes a serious estimation error because the time is delayed more than its true one. Thus, the NLOS nodes should be detected and eliminated for estimation. As a well-known NLOS detection scheme, an iterative minimum residual (IMR) scheme which has low calculation complexity is used for detection. However, the detection error exists in IMR scheme due to the measurement error. Therefore, in this paper, we propose a new IMR-based NLOS detection scheme and show its performance improvement by computer simulations.