Design of measurement-based correlation models for shadow fading

This paper deals with the design of measurement-based correlation models for shadow fading. Based on the correlation model, we design a simulation model using the sumof-sinusoids (SOS) method to enable the simulation of spatial lognormal processes characterizing real-world shadow fading scenarios. The model parameters of the simulation model are computed by applying the L
p
-norm method (LPNM). This method facilitates an excellent fitting of the simulation model's autocorrelation function (ACF) to that of measured channels. Our study includes an evaluation of all important statistical quantities of the proposed measurement-based simulation model, such as the probability density function (PDF), spatial ACF, decorrelation and coherence distance, as well as the level-crossing rate (LCR) and the average duration of fades (ADF). A comparison with the Gudmundson correlation model shows that the developed measurement-based correlation model outperforms the former one by far in terms of the goodness of fit to the measured data. The proposed measurement-based simulation model allows to study the effects of long-term fading on the performance of mobile communication systems under real-world shadow fading conditions.