Geometrical constrained least squares estimation in wireless location systems

Received signal strength indication (RSSI) is used in most cases of the localization system for its advantages of simplicity, requiring no additional hardware and low-cost. However, due to the irregular fluctuation, there are certain measured RSSIs which do not follow the empirical path loss model, while estimating the location of the blind node. So it's quite difficult to achieve high location accuracy. In this paper, according to the geometrical constraint between each distance, we propose a geometrical constrained least square estimation (GC-LSE) algorithm. This algorithm can both deal with the large errors that are mainly caused by the environmental change, and small errors that come from the devices' difference and operation mistakes. By reducing these errors we can greatly improve the location performance. With the numerical results from practical experiments, GC-LSE algorithm is proved to have a substantial improvement on location accuracy.