Statistical tuning of walfisch-bertoni pathloss prediction model based on building and street geometry sensitivity parameters in built−up terrains

Abstract

One of the underlying difficulties with the application of a prediction pathloss model for any environment is that no two areas are identical in the composition of the buildings and terrain. A pathloss model developed by Walfisch-Bertoni considers the impact of rooftops and building heights by using diffraction to predict average signal strength loss at street level. However, the error between the average path loss predicted by the model and that observed in practice will be smallest when the propagation environment conforms closely with the models assumptions (i.e. those urban environments exhibiting minimal variation in the height and separation of buildings). In particular, any building height variations can be expected to cause a significant error in the model predictions. Firstly, in this paper, we demonstrate the sensitivity impact of varying certain parameters on the performance W/B model in the study environments. The goal is to give the user an idea of the value associated with altering parameters that are in fact adjustable and showing the significance of the unchangeable parameters in the loss calculations. Secondly, a new statistical tuning method is proposed by extending the technique of Walfisch- Bertoni model which is now valid for generalized conditions in the CDMA2000 signal propagation environments. The results showed that tuned pathloss data agree strongly with measured data in the different study locations.