Dynamic estimation of optimum path loss model in a RSS positioning system

The development of location techniques in urban and indoor environments by using measurements of received signal strength (RSS) has as a main drawback the fact that in such environments the value of RSS depends on many unpredictable and changeable factors, not only on the actual distance between mobile station (MS) and the different base stations (BSs). With the aim of solving these limitations of RSS-based location systems, in this paper we present a novel technique to find out which model best describes the different propagations of signals traveling between the MS and the different BSs. This method estimates accurately and dynamically, all the path loss exponents that model the signal attenuation in propagation paths to each BS, only by using the received signal strength measurements that are obtained in each moment. Therefore, our method does not need any previous knowledge about user position or kind of propagation environment. From the estimation of the different path loss exponents, we obtain much more accurate distance estimations than those obtained by using a generic and unique propagation model previously fixed. We present the results obtained by using the methods described in this paper, from real measurements taken in a GSM network as well as from computer simulations. By using the techniques presented in this paper we can see the great improvement in the precision on estimating distances from the MS and the different BSs. The results obtained from the real measurements support the simulations results and show how that improvement in the distance estimation makes the location much more accurate.