Characterization and identification of ultrawideband radio propagation channels

This dissertation is concerned with the propagation of ultrawideband (UWB) electromagnetic impulses in complex multipath environments. In the first part the distortion of a UWB signal as it propagates through frequency selective materials is examined. The potential degree of distortion is investigated through experimental measurements, then the effect of that distortion on UWB communication and position location applications is examined. The second part considers a common specular channel model for ultrawideband indoor multipath propagation that also has much in common with indoor propagation models for non-ultrawideband signals. A number of properties of the model that can aid in design and analysis of communication systems in multipath channels but that are not explicit model parameters, such as power delay profile and delay spread, are derived as functions of the parameters of the model. The final part of the dissertation investigates performing channel identification: having two radios use the multipath channel as a source of common randomness in order to agree on a secret key that can by used for data encryption. The information theoretic aspects of secret sharing are considered as they apply to UWB channel identification, and bounds derived on the maximum identifier length. The practical aspects of channel identification are discussed, including a design methodology for the optimal quantizer, and techniques of varying complexity for aiding identifier calculation by sending feedback information over a public channel without revealing any information about the secret key. An array of Monte Carlo simulation results are provided demonstrating the relative performance of various algorithms. Finally the algorithms are applied to some measured multipath data in the presence of timing error.