An Alternate Model for Uplink Interference in CDMA Systems with Power Control

Abstract—Accurately modeling the statistics of the uplink interference power is an important problem in the design and analysis of second and third generation code division multiple access cellular systems. In the uplink, the total interference varies not only because of wireless channel effects such as shadowing and path loss, but also because of the randomness in the number of interfering mobiles and their locations. Conventionally, it has been argued this can be well modeled as a Gaussian distribution, especially in the presence of power control. In this paper, we show that when all the sources of randomness are considered together, the uplink inter-cell interference power is better modeled by a lognormal distribution. Using the well established theory of Poisson point processes to model spatial mobile distributions, we extend the moment matching or Fenton-Wilkinson method to determine the parameters of the approximating lognormal. Our results show that our lognormal approximation is several orders of magnitude more accurate than the conventional Gaussian approximation. These results have applications in cell planning and layout and, in general, in cellular system design and analysis. I