Modeling Received Signal Strength measurements for cellular network based positioning

This paper introduces a novel approach to model Received Signal Strength (RSS) measurements in cellular networks for user positioning needs. The RSS measurements are simulated by constructing a synthetic statistical cellular network, based on empirical data collected from a real life network. These statistics include conventional path loss model parameters, shadowing phenomenon including spatial correlation, and probabilities describing how many cell identities are measured at a time. The performance of user terminal positioning in the synthetic model is compared with real life measurement scenario by using a fingerprinting based K-nearest neighbor algorithm. It is shown that the obtained position error distributions match well with each other. The main advantage of the introduced network design is the possibility to study the performance of various position algorithms without requiring extensive measurement campaigns. In particular the model is useful in dimensioning different radio environment scenarios and support in preplanning of measurement campaigns. In addition, repeating the modeling process with different random values, it is possible to study uncommon occurrences in the system which would be difficult to reveal with limited real life measurement sets.