The communication systems community has been working towards integrated Software-Defined Radios (SDRs) and Cognitive Radios (CRs) that can reduce cost and enhance connectivity. In light of the technology bottleneck at the analog-to-digital converter (ADC) and the in-applicability of off-chip filters, the integrated analog front-end is entrusted with the task of sharp, linear, and programmable signal selection required for SDRs and CRs. Traditional analog filtering techniques, however, incur a high penalty in power consumption, area, and linearity to provide the required sharpness and programmability. Similarly, recent efforts that use compressive sensing to acquire wideband spectra have also faced a bottleneck in the complexity and re-configurability of the analog measurement front-end.

Towards enabling SDRs and CRs, this dissertation proposes a new perspective on the design of anti-alias filters that defies the traditional trade-off between cost, linearity, and programmability. The technique, termed Filtering by Aliasing (FA), anticipates the aliasing operation at the sampler instead of avoiding it. The pre-sampling circuitry is modulated, using the high-speed switching techniques popular in state-of-the-art receivers, to provide significantly enhanced filtering responses at the sampling instances. The dissertation describes how the FA technique, by varying the resistor of a single-pole passive RC filter for example, provides programmable anti-alias filtering comparable to a 7th-order Butterworth filter.

On the compressive sensing front, this dissertation proposes a new approach to the acquisition of sparse spectra using Random Filtering by Aliasing (RFA). RFA acknowledges the existence of noise in realistic spectra and accordingly simplifies the analog measurement stage, moving most of the complexity to the low-cost, highly reconfigurable digital domain. RFA achieves significantly better resolution, lower cost, and better programmability than existing schemes.

Energy consumption of the modern wireless communication systems is rapidly growing due to the ever-increasing data demand and the advanced solutions employed in order to address this demand, such as multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) techniques. These MIMO systems are power hungry, however, they are capable of changing the transmission parameters, such as number of spatial streams, number of transmitter/receiver antennas, modulation, code rate, and transmit power. They can thus choose the best mode out of possibly thousands of modes in order to optimize an objective function. This problem is referred to as the link adaptation problem.

In this work, we focus on the link adaptation for energy efficiency maximization problem, which is defined as choosing the optimal transmission mode to maximize the number of successfully transmitted bits per unit energy consumed by the link. We model the energy consumption and throughput performances of a MIMO-OFDM link and develop a practical link adaptation protocol, which senses the channel conditions and changes its transmission mode in real-time. It turns out that the brute force search, which is usually assumed in previous works, is prohibitively complex, especially when there are large numbers of transmit power levels to choose from. We analyze the relationship between the energy efficiency and transmit power, and prove that energy efficiency of a link is a single-peaked quasiconcave function of transmit power. This leads us to develop a low-complexity algorithm that finds a near-optimal transmit power and take this dimension out of the search space. We further prune the search space by analyzing the singular value decomposition of the channel and excluding the modes that use higher number of spatial streams than the channel can support. These algorithms and our novel formulations provide simpler computations and limit the search space into a much smaller set; hence reducing the computational complexity by orders of magnitude without sacrificing the performance.

The result of this work is a highly practical link adaptation protocol for maximizing the energy efficiency of modern wireless communication systems. Simulation results show orders of magnitude gain in the energy efficiency of the link. We also implemented the link adaptation protocol on real-time MIMO-OFDM radios and we report on the experimental results. To the best of our knowledge, this is the first reported testbed that is capable of performing energy-efficient fast link adaptation using PHY layer information.

With the continuous scaling of CMOS technology, two opposite trends can be observed. First trend is exponential improvement of the digital logic and second trend is linear degradation of the analog and RF components. Traditional methodology for designing analog and RF focuses on optimizing the performance of individual blocks such as analog-to-digital converters, frequency synthesizers, etc. Unfortunately, such traditional approach is limited by the unforgiving power-consumption-performance tradeoffs encountered in the design of each circuit block. As a result, high power consumption of RF and analog has become a major bottleneck for implementation of new communication algorithms.

Alternatively, one can "assist" the underperforming RF and analog blocks using cross-domain (RF-analog-digital) techniques. Such cross-domain approach leads to a significant saving in the power consumption while achieving the desired system performance (e.g. bit-error rate).

This dissertation explores cross-domain techniques to assist the analog-to digital converter, frequency synthesizer and antenna arrays. Novel cross-domain techniques are proposed to optimize the usage of dynamic range of analog-to-digital converter, to track and correct phase noise of the frequency synthesizer and to increase the processing bandwidth of phased array antennas. The proposed techniques have either been directly verified in experiment or as a minimum their implementation aspects have been fully considered in order to close the bridge between theoretical studies and practice.

By applying the proposed techniques to the system design of a 60 GHz radio link, we could improve the bit error rate of the link by several orders of magnitude while consuming an order of magnitude less power than the power that would otherwise have been consumed using traditional approach.

In order to introduce the benefits of Multiple-Input/Multiple-Output (MIMO) wireless solutions into the airborne environment for maximal effect, the airborne channel must be fully understood. While there have been theoretical models proposed for the airborne channel, there has been very little work toward providing a practical channel model which has been validated by actual an airborne platform. This work presents a characterization of practical performance gains of a MIMO system over a conventional SISO, in a mobile air-to-ground environment. Field measurements were collected with an airborne 4x4 MIMO-OFDM channel-sounding platform at altitudes, speeds and flight patterns approximating medium-endurance vehicles flying over various terrain. Ground stations placed in multiple locations (different scattering scenarios) measured channel responses in addition to actual throughput statistics. Our studies indicate that significant throughput and range gains are achievable with MIMO. We also show that depending on application requirements, these MIMO-enabled gains can be converted into considerable power savings. We also present a study of the effects of introducing MIMO-enabled signaling techniques (such as eigen beamforming and spatial multiplexing) on the total link-capacity of a system of uncoordinated, air-to-ground link-pairs deployed to a single area of operations. Captured channel measurements from the earlier real-world airborne study were inserted into our multi-link simulation environment. Trials were run under several representative aerial-deployment scenarios, revealing significant gains in link capacity. Finally, we consider the potential throughput enhancement delivered by full-duplex signaling and its limitations due to desensitization of receiver hardware by self-generated interference (SI). Existing SI cancellation solutions are prohibitive for long-range/airborne applications due to power handling limitations. They are also not easily scalable for an arbitrary number of MIMO antennas in arbitrary positions. A host-agnostic, high-power, adaptive SI canceler design is proposed and a hardware prototype is presented. Performance enhancement with an off-the-shelf host radio was demonstrated in the presence of varying SI signal profiles.

The performance of a packet-based communication system is determined by its packet error rate (PER) rather than its bit error rate (BER). Although the PER and the BER are correlated, it is not straightforward to calculate the PER using its BER due to the bit errors' correlation brought about by the channel coding. Therefore, the existing PER predictions are mostly heuristic methods.

The heuristic PER prediction methods, such as exponential effective SNR mapping (EESM) and mean mutual information per bit (MMIB), follow two steps: 1) compute an averaged single metric from the channel coefficients, and 2) use a simulated curve to map from the single metric to the PER. These methods require off-line simulations, curve-fitting, and parameter calibrations for every combination of modulation, code rate, and packet length. These requirements lower their applicability as modern systems support numerous modes (combinations of the above) thus needing an impractical number of off-line simulations. In this dissertation, we develop an analytical PER prediction method that requires no off-line simulations but delivers high accuracy for a wide range of transmission schemes and environments.

First, we assume that the system uses a uniform random interleaver and treat the bits' log-likelihood ratios (LLRs) as i.i.d. random variables. The LLR distribution is modeled as a mixture of folded normal distributions and approximated by a mixture of normal distributions. The PER is then calculated using a Gaussian Q-function approximation and the transfer function of the punctured convolutional code. Second, we focus on a realistic multiple-input multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) system that uses a repeated-pattern interleaver. Therefore, the LLRs are dependent and their correlation is governed by their locations in the packet. In general, there are three types of correlations: a) correlations across subcarriers, b) correlations across spatial streams, and c) correlations within a subcarrier. All the three types of correlations are essential and their effect is characterized in our PER prediction. The simulation result shows that the analytical method achieves higher prediction accuracy than all the existing heuristic methods.

The analytical PER prediction can be applied to the design of cyclic redundancy check (CRC) code, which is used for error detection. We consider a system employing a CRC code concatenated with a convolutional code. The undetected error probability of this system can be calculated by two methods: the exclusion method and the construction method. The exclusion method enumerates the error patterns of the convolutional code and tests if each of them is detectable. The construction method reduces complexity significantly by exploiting the equivalence of the undetected error probability to the frame error rate of an equivalent catastrophic convolutional code. We further propose a design of the CRC code for a specific convolutional code and codeword length such that the undetected error probability is minimized. This probability can be minimized when the CRC code detects the most probable residual error patterns at the output of the Viterbi decoder. In our example, the designed CRC codes have significant reduction in undetected error probability compared to the existing CRC codes.

Second order statistic based Spatial Interference Suppression has been shown effective in suppressing strong interference from jammers or concurrent transmissions. However, hardware impairments in the RF front end can be a major impediment to the design of such systems since the interference can be several orders of magnitude stronger than the signal. IQ mismatch is found to limit the interference suppression ability due to the creation of additional interference eigen modes from the image band of the interference source. This thesis shows how the concept of expanding the received signal sub-space can be used to mitigate the effect of IQ mismatch. Architectures for interference suppression in the presence of IQ mismatch are presented for both narrowband and wideband systems. Simulation results indicate 12 - 15 dB more interference suppression when using the proposed architecture to compensate for IQ mismatch in indoor environments.

With wireless communications becoming an essential part of human life, wireless technology advances to meet the increasing demands. New standards are introduced every couple of years to regulate the implementation of wireless systems. Most of modern standards are based on MIMO and OFDM signaling, which makes any time saving in a MIMO-OFDM receiver design cycle essential and the support of multi-standards in the same device highly desirable.

This work introduces a hardware implementation for a MIMO decoder accelerator, which is a software-programmable device that specializes in MIMO decoding, and MIMO signal processing in general, for OFDM systems. A VLSI implementation of the accelerator is introduced highlighting some of the implementation decisions and techniques to minimize the overall energy consumption of the accelerator hardware. The accelerator chip core area is 2.48mm2 in 65nm CMOS technology. Its average power consumption is 224.3 at 166MHz clock frequency. A deeply pipelined design for a powerful processing core allows the accelerator to achieve energy consumption figures competing with specialized designs. A single accelerator chip can be programmed to complete 4x4 QR decomposition, 4x4 Singular-Value Decomposition (SVD), 2x2 MMSE MIMO decoding, 4x4 MMSE MIMO decoding, or many other possible applications.

A simple design flow is presented to assist a MIMO-accelerator user in mapping a MIMO-related algorithm to a successful accelerator-based hardware implementation in no time. The accelerator, with its diversity and energy efficiency, can empower a wireless MIMO-OFDM receiver giving it an unparalleled advantage over regular fixed-data-path systems.

In the recent decades, there have been dramatic changes in the ways people create and share information, hence a demand for wireless communication systems with over 100 Gb/s data rate has been envisioned. Promising techniques have been devised by researchers to fulfill this expectation, such as interference alignment (IA) and THz communication and many more. In theory, given ideal assumptions, they show extraordinary increases in the promised data rate relative to current wireless communication systems. Nevertheless, because of the hardware impairments, many practical challenges have to be addressed before bringing these advanced techniques from theory to practice. In this dissertation, a selection of practical issues of interference alignment and THz communication are discussed.

First, we consider the synchronization and channel estimation problem associated with enabling IA. Before achieving IA, in view of the uncoordinated interference, a concurrent synchronization and channel estimation mechanism with robustness against co-channel interference is needed. We start by deriving the maximum-likelihood expressions for optimal estimators. Then, it is realized that such estimators are infeasible in practice for requiring a computationally costly 2-D search. To reduce the complexity, we propose a Zadoff-Chu sequence based pilot structure. By utilizing its unique properties, the 2-D search problem can be converted into two 1-D search problems. Experimental results verify the great performance of our proposed method in terms of both the residual frequency offset and the sum rate of an IA network. Second, we focus on the channel reciprocity in time division duplex (TDD) networks. The physical channel is reciprocal, but the channel considered by IA is not, because it also involves the radio frequency (RF) front-ends of the transceiver. The difference between the two individual circuits yields mismatch between the forward and backward channels. We provide the modeling of the non-reciprocity for the TDD MIMO channels, and propose a corresponding calibration method to minimize the mismatch. Experimental validation has been conducted. The results highlight the indispensability of the reciprocity calibration for approaching the network capacity promised by IA. Last, with a view to assessing THz communication system design requirements in practice, we measure the channels for both 350 GHz and 650 GHz carrier frequencies in a typical indoor environment. To obtain the spatial and temporal characteristics of the channel, the power angle profile (PAP) and the power delay profile (PDP) are measured using our proposed measurement systems and methods. From the measurement results, we have observed important facts regarding the THz wave propagation, which significantly affect the requirements for designing an indoor THz communication system.