Search

In this paper, we consider the problem of sensor selection for parameter estimation with correlated measurement noise. We seek optimal sensor activations by formulating an optimization problem, in which the estimation error, given by the trace of the inverse of the Bayesian Fisher information matrix, is minimized subject to energy constraints. Fisher information has been widely used as an effective sensor selection criterion. However, existing information-based sensor selection methods are limited to the case of uncorrelated noise or weakly correlated noise due to the use of approximate metrics. By contrast, here we derive the closed form of the Fisher information matrix with respect to sensor selection variables that is valid for any arbitrary noise correlation regime, and develop both a convex relaxation approach and a greedy algorithm to find near-optimal solutions. We further extend our framework of sensor selection to solve the problem of sensor scheduling, where a greedy algorithm is proposed to determine non-myopic (multi-time step ahead) sensor schedules. Lastly, numerical results are provided to illustrate the effectiveness of our approach, and to reveal the effect of noise correlation on estimation performance., IEEE Transactions on Signal Processing (accepted)

Wireless networks have revolutionized nowadays world by providing real time cost-efficient service and connectivity. Even such an unprecedented level of service could not fulfill the insatiable desire of the modern world for more advanced technologies. As a result, a great deal of attention has been directed towards (mobile) wireless sensor networks (WSNs) which are comprised of considerably cheap nodes that can cooperate to perform complex tasks in a distributed fashion in extremely harsh environments. Unique features of wireless environments, added complexity owing to mobility, distributed nature of the network setup, and tight performance and energy constraints, pose a challenge for researchers to devise systems which strike a proper balance between performance and resource utilization. We study some of the fundamental challenges of wireless (sensor) networks associated with resource efficiency, scalability, and location awareness. The pivotal point which distinguishes our studies from existing literature is employing the concept of sparse reconstruction and compressive sensing (CS) in our problem formulation and system design. We explore sparse structures embedded within the models we deal with and try to benefit from the undersampling offered by incorporating sparsity and thereby developing sparsity-aware system-level solutions. We prove that looking at these challenges from our perspective not only guarantees an expected cost efficiency due to taking less measurements, but also if properly designed, can promise an acceptable accuracy.

This thesis is dedicated to transceiver designs for high data-rate wireless communication systems with rapidly moving terminals. The challenges are two-fold. On the one hand, more spectral bandwidth of the transmitted signals is required by future wireless systems to obtain higher transmission rates, which can result in the frequency selectivity of the communication channels. On the other hand, Doppler effects emerge when high mobile speeds are present, which can result in the time selectivity of the communication channels. Therefore, it is likely that future wireless communication systems operate in doubly-selective channels, which impose many difficulties on transceiver designs. In this thesis, we investigate these challenges in the following four scenarios, and propose a number of corresponding solutions. OFDM over Narrowband Channels: Orthogonal frequency-divisionmultiplexing (OFDM) is a typical multiple-carrier transmission technique. In a narrowband scenario, Doppler effects are well approximated as frequency shifts. In this manner, a narrowband doubly-selective channel for OFDM systems can be approximately characterized as a banded matrix especially when a basis expansion model (BEM) is exploited to model the channel. It thus allows to reduce the complexity of the channel equalization. However, there are various different BEM’s available. We identify a particular BEM which leads to a more efficient hardware architecture than other choices, while still maintaining a high modeling accuracy. OFDM over Wideband Channels: The Doppler effect manifests itself as a distinct phenomenon in wideband channels compared to narrowband channels. Specifically, the wideband signal waveformismeasurably dilated or compressed when Doppler is present rather than just frequency-shifted. This unique nature ofwideband time-varying channels requires new designs for wideband OFDM systems. We first quantify the amount of interference resulting from wideband doubly-selective channels which follow the multi-scale/multi-lag (MSML) model. Then we discuss an equalization method for wideband channels either in the frequency domain or in the time domain. A novel optimum resampling procedure is also introduced, which is normally unnecessary in narrowband systems. Multi-Rate Transmissions over Wideband Channels: Traditional multi-carrier transmission schemes, e.g., OFDM, use a uniformdata rate on each subcarrier, which is inherently mismatched with wideband time-varying channels. In fact, the time variation of wideband channels, i.e., the Doppler scales, imply a non-uniform sampling mechanism. To mitigate this, we propose a novel multi-rate transmission scheme by placing the information symbols at different non-overlapping sub-bands where each sub-band has a distinctive bandwidth. To combat the MSML effect of the channel, a filterbank is deployed at the receiver, where each branch of the filterbank samples the received signal at a corresponding rate. By selecting a proper transmit/receiver pulse, the effective input/output relationship can be captured by a block-diagonal channel, with each diagonal block being a banded matrix similarly as seen in narrowband OFDM systems. The benefit of this similarity is that existing low-complexity equalizers can be adopted for wideband communications. Robust Multi-band Transmissions over Wideband Channels: Accurate channel estimation for wideband doubly-selective channels is challenging and troublesome. Adaptive channel equalization is thus attractive since it does not require precise channel information and is robust to various prevailing environmental conditions. When the MSML effect emerges in wideband channels, it is not wise to adopt existing adaptive equalization designs that are previously used in other scenarios, e.g., narrowband channels. We adopt a multi-band frequency-division multiplexing (FDM) signal waveform at the transmitter to reduce the equalization complexity, while maintaining a high data rate. By carefully designing the transmit pulse, our proposed multi-layer turbo equalization, using a phase-locked loop (PLL) followed by a time-invariant finite impulse response (FIR) filter, is capable of equalizing such MSML channels.

The great demand for location-aware wireless sensor networks (WSNs) motivates the research in this thesis. The unique characteristics of WSNs impose numerous challenges on localization and communication. In this thesis, we handle some key challenges and provide affordable solutions. Impulse radio ultra wideband (IR-UWB) is employed as the fundamental technology for both localization and communication due to its distinctive advantages in accurate ranging and reliable communication. The following aspects are treated in this thesis. Transmitted-reference (TR) UWB communication systems: IR-UWB processing in the digital domain usually asks for very high sampling rates. The TR-UWB scheme allows for sub-Nyquist rate sampling by correlating the received pulse sequence with its delayed version in the analog domain. Thus, it avoids the daunting Nyquist sampling rate, relaxes the stringent synchronization requirements, and only asks for aggregate channel coefficients. A data model including various kinds of interferences is employed, and then a complete receiver is proposed including signal detection, channel estimation, synchronization and equalization. Theoretical ranging bounds and practical ranging methods based on IR-UWB: We investigate the theoretical ranging accuracy of a novel method, which exploits the range information in both the amplitude and the time delay of the received signal. The investigations are conducted not only for an additive white Gaussian noise (AWGN) channel with attenuation, but also for an AWGN channel with both attenuation and shadowing. Furthermore, a practical ranging method based on time-of-arrival (TOA) estimation using UWB IRs is developed. Stroboscopic sampling is employed to sacrifice transmission efficiency for a lower sampling rate. Moreover, it can maintain the same ranging resolution as Nyquist sampling can achieve. Due to the long preamble required by stroboscopic sampling, the clock drift, which is an accumulative effect over time caused by the relative clock skew between different clocks, is one of the main error sources in TOA estimation. Therefore, TOA estimation methods with clock drift calibration are explored to dramatically mitigate the influence of the drift. Various localization and tracking methods: Extended multi-dimensional scaling (MDS): Since the classical MDS cannot be applied to general networks with missing links, we extend the classical MDS algorithm to deal with a special kind of network with specific missing links. Our goal is to jointly estimate the positions of all the nodes given partial pairwise distance measurements up to a translation, rotation, and reflection. Reference-free time-based localization: Low-complexity least-squares (LS) estimators based on time-of-arrival (TOA) or time-difference-of-arrival (TDOA) measurements have been developed in literature to locate a target node with the help of anchors (nodes with known positions). They require to select a reference anchor in order to cancel nuisance parameters or relax stringent synchronization requirements, and suffer from a poor reference selection. We propose reference-free localization estimators based on TOA measurements to decouple the reference dependency. Furthermore, we generalize existing reference-based closed-form localization estimators using TOA or TDOA measurements, and shed new light on their relations to clarify some confusions that still persist in recent literature. Robust time-based localization: Time-based localization approaches attract a lot of interest due to their high accuracy and potentially low cost for WSNs. However, time-based localization is tightly coupled with clock synchronization. Thus, the reliability of timestamps in time-based localization becomes an important yet challenging task to deal with. Regardless of the reliability of the timestamps from the target node, we propose a novel ranging protocol, namely asymmetric trip ranging (ATR), which leads to localization methods that are naturally immune to internal attacks mounted by a compromised target node. Robust localization strategies using the ATR protocol based on TOA measurements are proposed to localize a target node with the help of anchors for asynchronous networks. Kalman tracking: Due to the nonlinearity of the localization problem, a Kalman filter (KF) is usually replaced by an extended KF (EKF) for tracking a mobile target. However, the modeling errors inherently contained in the EKF degrade the tracking performance. Therefore, we make use of the ATR protocol again, carry out exact linearizations, and achieve a KF based on a linear measurement model to track a mobile target with the aid of fixed anchors.