Communication and Localization in UWB Sensor Networks: A Synergetic Approach
ReferencePhD Thesis, Logos Verlag Berlin, 2013.
In this thesis, a novel sensor network paradigm is proposed and studied, inspired by the fusion of wireless communication, localization and imaging. Wireless sensor networks will open a fascinating world of ubiquitous and seamless connectivity not only between individuals but also between devices and objects in our daily life. The key to this vision is a universal low-power, low-complexity and low-cost transceiver unit that provides scalable data communication as well as location and environmental information. Ultra-Wideband (UWB) technology with its rich design space can meet the challenging requirements of future wireless sensor networks. This is the consequence of a paradigm shift compared to narrowband communication: due to the huge bandwidth available, we can trade off bandwidth efficiency against other figures of merit. The major design criterion is not data rate anymore, but rather power consumption and hardware complexity. Within the group of hardware-aware system designs, UWB impulse radio with energy detection receivers are of particular relevance and well known for their efficient implementation. The contribution of this thesis is the comprehensive study of sensor networks with generalized energy detection receivers, where we focus on innovative and efficient approaches for communication and localization and their synergy.
The first part of this thesis develops a framework for location-aware optimization of data transmission with generalized energy detection receivers. This framework is based on a Signal-to-Interference-plus-Noise-Ratio expression. It covers receiver as well as transmitter optimization, where narrowband interference suppression is also taken into account. Conventional approaches attempt to adapt the transceiver directly to the channel state. They require the knowledge of the channel state over the full transmission bandwidth. Due to the huge bandwidth of UWB, the estimation and dissemination of channel state information requires high complexity and is very expensive and power hungry. To circumvent this problem, we propose adjusting the transceiver to the node position. This is done by modeling the channel impulse response as a random process with location dependent parameters, which can be estimated in an off-line training phase. The data transmission is then optimized based on the position of the node -- a more accessible type of information in UWB networks that may already be available. In the next step, we extend the optimization to multiuser transmission. This leads to an increase in the sum data rate, while maintaining the low complexity of the nodes. We conclude from performance evaluations that location information can improve the performance of low complexity and low power UWB communication.
The second part of this thesis is dedicated to localization. We focus on the estimation of the travel time of the radio signal, which is related to the distance and, thus, to the location of the node. Many existing approaches for time of arrival (ToA) estimation of UWB signals require high speed sampling of at least twice the bandwidth of the received signal. This leads to high power consumption and high hardware complexity. We propose to perform ToA estimation at the output of an energy detection receiver. This allows the sampling rates to be much lower keeping the complexity and the power consumption low. In order to achieve high performance, we derive the maximum likelihood timing estimator for the generalized energy detection receiver. We again model the channel impulse response as a random process and show that location-aware a-priori knowledge of its distribution can increase the performance significantly. Approximations of the maximum likelihood estimator lead to a family of low-complexity timing estimators that trade lower estimation accuracy for lower computing time. Additionally, we analyze spectral timing estimation at the energy detector output. We derive the accuracy in a multipath channel analytically, which provides an insight into the fundamental performance scaling of the estimation error in multipath.
So far, we show that location-aware a-priori channel knowledge is beneficial for communication as well as position estimation. However, this requires a database that maps positions to channel characteristics. The training phase to acquire this database can be difficult to implement, time-consuming and may require many training samples. The solution to this problem is imaging. In the third part of this thesis, we present radar imaging based channel prediction. Using this method, the channel characteristics can be obtained for every position in a stationary environment from just a few training samples. An extensive measurement campaign proves the practicality of the presented approach. Finally, we draw conclusions and give an outlook on future research in UWB communication, localization and imaging.
This publication is currently not available for download.