Dynamic properties of wireless channels with application to adaptive modulation and multiuser diversity systems

In this thesis we analyze the performance of antenna diversity systems, adaptive modulation, and multiuser diversity systems. We first study second-order statistical properties of various antenna diversity combining schemes in terms of the average outage duration and the average level crossing rate. These second-order statistical properties are important for system design issues such as the interleaver depth design and the packet length optimization. We focus initially on the effect of co-channel interference and minimal power constraint as well as fading correlation, channel unbalance, and speed of users. We then extend the applicability of the Rice formula to the scenarios when multiple independent random processes are combined in different forms. This allows us to characterize the second-order statistics of more complicated communication systems and networks of current interest. Applications included for example multi-hop communication systems with and without selection combining as well as communication networks with routing diversity.; The thesis then investigates the packet delay behavior in a wireless link employing adaptive modulation. Since the outage probability of adaptive modulation can be quite high, especially for channels with low average signal-to-noise ratio (SNR), adaptive systems require buffering of the input data. Because separate-layer consideration may not be appropriate for the time-varying wireless channels, this design challenge prompts a new design methodology that jointly optimizes the allocation of system resources through a cross-layer design. In this work, we conduct a cross-layer performance analysis of an adaptive multicarrier transmission system with a finite data buffer at the transmitter. We formulate such a system as an M/G/1 queue in which the departure rates form a discrete set and characterize the delay performance by the buffer overflow probability and the average packet waiting time.; Finally we study the performance tradeoffs in the multiuser diversity systems with affordable-rate transmission and adaptive transmission. We first characterize the performance criteria included achievable system capacity/system spectral efficiency as well as the degree of fairness, access probability, average access time, and average access rate of individual users. These performance are studied under the throughput-maximizing scheduling and the proportional fairness scheduling algorithms, which are critical for the viability of these scheduling algorithms and are useful to understand the capacity/fairness tradeoff of multiuser diversity scheduling schemes. Based on these studies, we propose and analyze new scheduling schemes to reduce the channel quality feedback load as well a more flexible balance/tradeoff between the system achievable capacity and the fairness among users.