Adaptive receiver design and optimal resource allocation strategies for fading channels.

摘要

The mobile wireless environment has been a challenge to reliable communications because of the time-varying nature of the channel. Detrimental effects such as path loss, shadowing, and multipath fading can greatly attenuate the transmitted signal. Therefore, adaptive channel estimation and data detecting algorithms must be designed for such channels. Moreover, in a multi-user system, dynamic resource allocation is an important means to transmit information efficiently through the varying channel.; In this thesis we first propose two adaptive feedback maximum-likelihood detection techniques, a decision-feedback decoder and an output-feedback decoder, for coded signals transmitted over channels with correlated fading. Both analysis and simulation results demonstrate that they have far better performance than the conventional decoder. We also propose a simple improvement to conventional decoders by using a weighted metric. The BER performance of all these decoders is analyzed through a sliding window decoding method.; Next we derive the ergodic (Shannon) capacity region and optimal dynamic resource allocation for an M-user fading broadcast channel under code-division with and without successive decoding, time-division, and frequency-division. For this channel we also derive the outage and zero-outage capacity regions and the corresponding optimal resource allocation strategies under different spectrum-sharing techniques. We obtain the outage capacity region implicitly by deriving the minimum common outage probability or the outage probability region for a given rate vector. The corresponding optimal power allocation scheme is a multi-user generalization of the single-user threshold-decision rule.; Finally, we obtain the outage capacity region and optimal power allocation for fading multiple access channels. Successive decoding is proved to be optimal and iterative algorithms are proposed to obtain the optimal decoding order and power allocation in each fading state under the average power and outage probability constraints of each user. We also obtain the average power regions that can support a rate vector with the given average outage probability of each user satisfied.