Signal strength prediction in built-up areas. Part 2: Signal variability

Analysis of mobile radio propagation results in London at 168, 455 and 900 MHz has shown that over small areas the signal envelope is well described by Rayleigh statistics, the local mean over larger similar areas being lognormally distributed. A comparison of different statistical distributions has been undertaken to provide a bridge between the local and global statistics and it shows that the experimental data are best fitted by a Suzuki (Rayleigh plus log-normal) distribution, the lognormal part having a standard deviation of approximately 4 dB. Values near the tails of the distribution can thereby be described with good accuracy, and hence the propagation prediction model is useful for area coverage estimation and frequency reuse planning.