Abstract
Ultra-wideband (UWB) communications exhibit many special characteristics, one of which is the random clustering phenomenon in the resolved channel multipaths. The randomness in path arrivals will, however, make exact performance analysis quite challenging. In this thesis, an analytical framework for precise characterization of a unified UWB channel structure is developed. The channel structure under consideration can comprise both the IEEE 802.15.3a and IEEE 802.15.4a channel models of great practical interests. The key idea is to capture the average eRect of the cluster and ray arrival processes through the extended density functions. The density functions are verified to be useful in deriving exact channel statistical quantities of interests, including the second-order and fourth-order joint moments. Precise performance analysis for UWB systems regarding practical channel and interference scenarios can therefore be carried out based on the exact channel analytical results. Specifically, the performance analysis and optimization for direct-sequence UWB (DS- UWB) and multi-band orthogonal frequency-division multiplexing (MB-OFDM) systems are investigated. For the DS-UWB systems, improved analysis that do not rely on simplified interference assumption and random spreading as in the previous related works is conducted. For the MB-OFDM systems, the discrete-time receiver operation, the pulse-shaping function, and the practical cyclic addition scheme that can achieve a significant performance gain are all included in the analysis. Numerical results are also provided to verify the precise performance evaluation and the practical value of system optimization by our analytical results. It is noticed that, due to the lack of good analytical results for the special channel characteristics, previous methodology for estimation of the channel parameters can cause inconvenience and ambiguity. As a novel application of our analytical results, a system- atic approach for estimating the UWB channel parameters is also developed in the thesis. The estimation is formulated as an optimization problem with an explicit analytical metric function that can be systematically solved. A limited set of good candidate parameters can rapidly be obtained, and further statistical tests with moderate eRorts can give the best estimation. The eRectiveness of the proposed method will be demonstrated by showing the achieved good agreements under the considered statistical profiles. With the unified analytical framework and the practical applications in the thesis, it is expected that our work can provide fundamental contributions to the research field of UWB communications.