Bit Error Probabilities of Ncfsk and Dpsk Signals in Microcellular Mobile Radio Systems

Introduction: To extend the present mediumjlarge cell mobile radio systems [I] towards ubiquitous communications, the microcellular approach looks promising [Z]. A major problem in cellular radio systems, regardless of having micro or mediumflarge cell architecture, is the presence of cochannel interference because of frequency reuse. Extensive studies on the effects of cochannel interference in mediumjlarge cell systems have been made [3-51. Models from Rayleigh fading [31, lognormal shadowing [4] to superimposed Rayleigh fading and lognormal shadowing [SI have been used for signal environments. There is a common assumption in all these studies, however, that all signals received, desired and undesired, have the same statistical characteristics. For example, in Reference 3, both the desired signal and the undesired cochannel interferers are assumed to be subject to Rayleigh fading. To evaluate the performance of microcellular mobile radio systems, a microcell interference model, in which different fading statistics are used to characterise the desired and undesired signals, is introduced in Reference 6. The desired signal is assumed to have Rician statistics implying that a dominant multipath reflection exists in within-cell transmission [7]. The interference signals from cochannel cells are assumed to he subject to Rayleigh fading because of the absence of a line-of-sight propagation. This is the Rician/ Rayleigh microcell interference model [SI. This model has been used to study the outage probability and spectral eniciency of microcellular mobile radio systems [6, E]. This Letter further investigates the Rician/Rayleigh interference model by deriving the probability density function (PDF) of the power ratio between the desired and interfering signals. The PDF is then used to ohtaio the bit error probabilities of noncoherent frequency shift-keying (NCFSK) and differential phase shift-keying (DPSK) signals in microcellular systems with cochannel interference. The hit error probability issue in microcellular radio systems has been studied in Reference 9. However, in Reference 9, the cochannel interferers were approximated by independent Gaussian random variables. The microcell cochannel interference is thus not characterised using its fading statistics. In this Letter, we will derive the hit error probabilities in microcellular radio systems by considering both the Rician fading characteristics of the desired signal and the Rayleigh fading characteristics of the interferers. In situations where both the desired signal and the interferers have Rayleigh fading characteristics, the results apply to mediumjlarge cell systems.