A geometry-based algorithm for TDoA-based localization in the presence of outliers

Localization of non-cooperative radio frequency transmitters in wireless networks can be performed by using Time Difference of Arrival (TDoA). However, TDoA measurements are often degraded by limited signal-to-noise ratio, multipath components, non-line-of-sight (NLOS) visibility, etc. This causes the TDoA measurements to have strong outliers, which cannot easily be modelled mathematically. This paper investigates algorithms to perform TDoA-based localization in the presence of outlier measurements. Our algorithm is based on a geometric interpretation, rather than a mathematical, noise model-based approach. We show that our algorithm outperforms existing algorithms when applied to real-world data in a vehicular scenario.