Ultra-wideband based communications and localization in wireless sensor networks

The key motivations for using ultra-wideband (UWB) communication systems are their capability for providing high capacity wireless communications, as well as the availability of technology to implement and generate UWB signals with relatively low complexity.
Non-coherent receivers with no channel estimation have been proposed to make the UWB technology attractive for applications where low cost and low power consumption are playing an important role, as in the case of wireless sensor networks (WSNs). Transmitted Reference (TR) signaling, in combination with an autocorrelation receiver (AcR) is especially
suitable in this context. The large bandwidth of UWB technology does not only provide the possibility to transmit at a very high data rate, but also provides very accurate temporal and spatial information that can be used for precise timing offset estimation.
UWB Impulse Radio (IR) uses sub-nanosecond pulses which provide a high resolution capability in the time domain, making it attractive for accurate wireless localization. Indeed, its ability to resolve multipath components makes it possible to obtain accurate location
estimates without the need for complex estimation algorithms. This facilitates many applications such as location-aware sensor networking.

This thesis presents several contributions towards developing UWB technology for short-range low to medium data rate WSN applications. First, different low complexity TR-UWB receivers are thoroughly analyzed, using the Gaussian approximation on the noise terms in the receiver statistics, and then using the equivalent system and noise models of these receivers. Both approaches lead to the bit error performance analysis, which is also presented in this thesis. The TR-UWB systems were modelled as polynomial nonlinear systems. Since the statistics of the radio channel are used in the system modelling as
well as in the performance analysis, appropriate channel models for UWB impulse radio systems are discussed in this thesis.
Further, synchronization issues are addressed. To this end, a novel data-aided timing acquisition technique for frame-level synchronization of DTR-UWB systems is suggested. It is based on incorporating parallel integration-and-dump circuits within pulse-pair correlator branches to improve considerably the energy capture in the presence of timing offset.
Moreover, a simple algorithm for fine synchronization of low complexity TR-UWB systems,assuming no inter-symbol interference, is proposed. It uses energy collected at the symbol rate, thus reducing considerably the implementation complexity. If the hardware clock timing error is known, timing offset estimation problems are analogical to ranging problems,and the proposed algorithm can readily be used as a new distance measurement technique.
It is shown, in this thesis, that the proposed ranging approach allows localization accuracy in the centimeter range using TR-UWB systems with data rates up to 5 Mb/s.