Software-Defined Architectures for Spectrally Efficient Cognitive Networking in Extreme Environments

The objective of this dissertation is the design, development, and experimental evaluation of novel algorithms and reconfigurable radio architectures for spectrally efficient cognitive networking in terrestrial, airborne, and underwater environments. Next-generation wireless communication architectures and networking protocols that maximize spectrum utilization efficiency in congested/contested or low-spectral availability (extreme) communication environments can enable a rich body of applications with unprecedented societal impact. In recent years, underwater wireless networks have attracted significant attention for military and commercial applications including oceanographic data collection, disaster prevention, tactical surveillance, offshore exploration, and pollution monitoring. Unmanned aerial systems that are autonomously networked and fully mobile can assist humans in extreme or difficult-to-reach environments and provide cost-effective wireless connectivity for devices without infrastructure coverage.;Cognitive radio (CR) has emerged as a promising technology to maximize spectral efficiency in dynamically changing communication environments by adaptively reconfiguring radio communication parameters. At the same time, the fast developing technology of software-defined radio (SDR) platforms has enabled hardware realization of cognitive radio algorithms for opportunistic spectrum access. However, existing algorithmic designs and protocols for shared spectrum access do not effectively capture the interdependencies between radio parameters at the physical (PHY), medium-access control (MAC), and network (NET) layers of the network protocol stack. In addition, existing off-the-shelf radio platforms and SDR programmable architectures are far from fulfilling runtime adaptation and reconfiguration across PHY, MAC, and NET layers. Spectrum allocation in cognitive networks with multi-hop communication requirements depends on the location, network traffic load, and interference profile at each network node. As a result, the development and implementation of algorithms and cross-layer reconfigurable radio platforms that can jointly treat space, time, and frequency as a unified resource to be dynamically optimized according to inter- and intra-network interference constraints is of fundamental importance.;In the next chapters, we present novel algorithmic and software/hardware implementation developments toward the deployment of spectrally efficient terrestrial, airborne, and underwater wireless networks. In Chapter 1 we review the state-of-art in commercially available SDR platforms, describe their software and hardware capabilities, and classify them based on their ability to enable rapid prototyping and advance experimental research in wireless networks. Chapter 2 discusses system design and implementation details toward real-time evaluation of a software-radio platform for all-spectrum cognitive channelization in the presence of narrowband or wideband primary stations. All-spectrum channelization is achieved by designing maximum signal-to-interference-plus-noise ratio (SINR) waveforms that span the whole continuum of the device-accessible spectrum, while satisfying peak power and interference temperature (IT) constraints for the secondary and primary users, respectively. In Chapter 3, we introduce the concept of all-spectrum channelization based on max-SINR optimized sparse-binary waveforms, we propose optimal and suboptimal waveform design algorithms, and evaluate their SINR and bit-error-rate (BER) performance in an SDR testbed. Chapter 4 considers the problem of channel estimation with minimal pilot signaling in multi-cell multi-user multi-input multi-output (MIMO) systems with very large antenna arrays at the base station, and proposes a least-squares (LS)-type algorithm that iteratively extracts channel and data estimates from a short record of data measurements. Our algorithmic developments toward spectrally-efficient cognitive networking through joint optimization of channel access code-waveforms and routes in a multi-hop network are described in Chapter 5. Algorithmic designs are software optimized on heterogeneous multi-core general-purpose processor (GPP)-based SDR architectures by leveraging a novel software-radio framework that offers self-optimization and real-time adaptation capabilities at the PHY, MAC, and NET layers of the network protocol stack. Our system design approach is experimentally validated under realistic conditions in a large-scale hybrid ground-air testbed deployment. Chapter 6 reviews the state-of-art in software and hardware platforms for underwater wireless networking and proposes a software-defined acoustic modem prototype that enables (i) cognitive reconfiguration of PHY/MAC parameters, and (ii) cross-technology communication adaptation. The proposed modem design is evaluated in terms of effective communication data rate in both water tank and lake testbed setups. In Chapter 7, we present a novel receiver configuration for code-waveform-based multiple-access underwater communications. The proposed receiver is fully reconfigurable and executes (i) all-spectrum cognitive channelization, and (ii) combined synchronization, channel estimation, and demodulation. Experimental evaluation in terms of SINR and BER show that all-spectrum channelization is a powerful proposition for underwater communications. At the same time, the proposed receiver design can significantly enhance bandwidth utilization. Finally, in Chapter 8, we focus on challenging practical issues that arise in underwater acoustic sensor network setups where co-located multi-antenna sensor deployment is not feasible due to power, computation, and hardware limitations, and design, implement, and evaluate an underwater receiver structure that accounts for multiple carrier frequency and timing offsets in virtual (distributed) MIMO underwater systems.