Analysis of multiuser diversity in time-varying channels

Research in recent years has been focusing on the role of multiuser diversity in enhancing the throughput of wireless networks. Channel fading is now seen as a source of randomization that can be exploited by a smart scheduler with partial channel information, for example, in terms of data rates requested by the users. In this paper, we analyze the performance of a multiuser diversity system, taking into account the feedback errors due to channel variability. Channel estimation errors are modeled as Gaussian random variables with variance depending on the Doppler spread. Based on a block fading channel model, we present an expression for the throughput, both as a function of the key system parameters, such as packet length and data rate thresholds, and channel characteristics, such as Doppler spread. From the analysis of the performance, we present a tradeoff between multiuser diversity and mobility, and also between rate adaptivity and channel estimation errors. It is shown that multiuser diversity gain is reduced for increasing receiver mobility, and that the throughput loss depends also on the distribution of the data rate thresholds. Finally, we use the expressions derived in the paper in order to adaptively reduce the loss due to feedback errors. Given a set of possible data rates, the users can choose the most suitable one based not only on the set of thresholds and signal strength, but also on mobility conditions and system parameters, such as packet length.