A spatio-temporal model for indoor non-line-of-sight (NLOS) impulse-based ultra-wideband channels, comprising clusters of multipath components with different times and angles of arrival, is proposed. This model is a spatial extension of the fixed-cluster temporal channel model, based on clustering patterns observed both in the times and angles of arrival of multipath components. It is demonstrated that the proposed extension accurately models the spatial correlation as well as the temporal properties of measured ultra-wideband signals in an indoor NLOS environment.

In this paper, a narrowband interference (NBI) mitigation scheme for ultra-wide bandwidth (UWB) signals using multiple receive antennas is examined. The low spatial fading of UWB signals relative to NBI signals is exploited to provide "interference selection diversity". Whereas classical selection diversity is designed to maximize the desired received signal power, the aim of interference selection diversity is to minimize the effective NBI power. Expressions for the probability of error of the selection diversity scheme are derived for both Rayleigh and Ricean NBI fading scenarios. The improvement in received signal-to-interference ratio is also mathematically investigated. It is shown that doubling the number of antennas results in a 3-dB performance improvement for the Rayleigh fading case. Less substantial gains are observed under Ricean fading.

Traditional synchronization techniques applied to impulse-radio ultra-wideband (UWB) result in prohibitively long acquisition times, due to the extremely large search space. Additionally, in dense multipath environments, there exist a larger number of cells within the uncertainty region that can lead to acquisition lock. Locking to an arbitrary multipath component may result in unacceptable performance for many applications (range error in positioning systems for example). In this paper, we present a modified framework for the analysis of UWB acquisition which accommodates multiple lock cells. The framework divides the acquisition process into two distinct phases. The two phases are termed "coarse" and "fine" acquisition. The coarse acquisition phase is a fast implementation of traditional serial search which takes advantage of the large number of cells which can terminate the search process. Fine acquisition exploits statistics derived from the first phase and the clustered nature of multipath arrivals to determine the earliest arriving path, even when it is severely attenuated. We show that the first phase provides a substantial improvement in mean acquisition time compared with traditional serial search and that the second phase provides robust estimation of the first arriving path in ranging applications.