Dynamic Clustering Approach to Compress Radio Database for Efficient Localization in Fingerprinting Systems

Radio fingerprinting is the process which identifies a cellular phone / other radio transmitter by the unique "fingerprint" which characterizes the signal transmission. Location fingerprinting determines receiver position with the help of database of RSS measurements from different sources .Location fingerprinting a positioning method that explores existing infrastructures e.g. cellular networks or WLANs. Dynamic Clustering technique to compress the radio database in the context of cellular fingerprinting systems has been proposed. The aim of the dynamic clustering technique is to reduce transmission load and computation cost in the mobile-based implementations. The presented method may be called Dynamic Block-based Weighted Clustering (DBWC) technique, which overcomes the issue of BWC that is it requires fix number of cluster which is not feasible in dynamic environment. DBWC is applied in a location-radio signal space, and assigns different weight factors to the location and radio components. The main idea of this work is to propose a dynamic clustering technique well-tailored to the structure of the radio database in a fingerprinting system. Here, we tend to outline the idea of feature types in association with database records. The obtained result confirms the efficiency of DBWC.