Cellular Network Fingerprint Localization Simulation: A Soft Computing Approach

Cellular Network Localization (CNL) is a vital source of information to a wide range of emerging applications, to name a few, Location based services, traffic analysis and sensing, vehicles´ anti-theft systems. Developing, validating, and testing of such applications require massive data sets of CNL - a prerequisite that is impractical to be collected. Furthermore, it can not be simulated using hard computing models due to the lack of estimators that can ascribe accurate level of localization accuracy as to the change in various spatial properties. In this paper, we propose a soft computing model to simulate the CNL performance, called SCMCL. This model carries out the estimation of CNL performance according to a knowledge base built on the basis of empirical findings reported in the literature. SCMCL is fed with real-life vehicular traces and produces CNL. SCMCL is validated and demonstrated throughout various operation conditions.