Statistical properties of loaded wireless multicarrier systems

Abstract

Multicarrier modulation has established itself as an appealing option for high-performance wireless communication systems. When the channel varies slowly in time, such as in applications with low terminal mobility, adaptive loading further enhances the system performance considerably. However, even with adaptive loading, significant performance fluctuations still occur, which impact the higher layers and the overall system behavior. First, a good understanding of the nature of these fluctuations is crucial to design better higher layers that are able to tolerate or combat this variability. For this reason, we develop a full statistical description of a loaded wireless multicarrier system that accurately characterizes its behavior. Our model abstracts the loaded system as an equivalent single carrier system with flat log-normal fading and has the channel tap as a single parameter. Second, with our model, the physical layer can now be incorporated in an abstracted fashion during simulation, resulting in significant speed-ups. Indeed, the alternative would be to include the entire loaded multicarrier system in these simulations, which is highly time consuming. Third, our statistical model is also important for the mathematical analysis of protocols, which requires knowledge of the physical layer statistics. We derive such statistics for both correlated and uncorrelated Rayleigh fading.