The introduction of Machine Learning (ML) and Deep Learning (DL) techniques into modern radio communications system, a field known as Radio Frequency Machine Learning (RFML), has the potential to provide increased performance and flexibility when compared to traditional signal processing techniques and has broad utility in both the commercial and defense sectors. Existing RFML systems predominately utilize supervised learning solutions in which the training process is performed offline, before deployment, and the learned model remains fixed once deployed. The inflexibility of these systems means that, while they are appropriate for the conditions assumed during offline training, they show limited adaptability to changes in the propagation environment and transmitter/receiver hardware, leading to significant performance degradation. Given the fluidity of modern communication environments, this rigidness has limited the widespread adoption of RFML solutions to date. Transfer Learning (TL) is a means to mitigate such performance degradations by re-using prior knowledge learned from a source domain and task to improve performance on a "similar" target domain and task. However, the benefits of TL have yet to be fully demonstrated and integrated into RFML systems. This dissertation begins by clearly defining the problem space of RF TL through a domain-specific TL taxonomy for RFML that provides common language and terminology with concrete and Radio Frequency (RF)-specific example use- cases. Then, the impacts of the RF domain, characterized by the hardware and channel environment(s), and task, characterized by the application(s) being addressed, on performance are studied, and methods and metrics for predicting and quantifying RF TL performance are examined. In total, this work provides the foundational knowledge to more reliably use TL approaches in RF contexts and opens directions for future work that will improve the robustness and increase the deployability of RFML. / Doctor of Philosophy / The field of Radio Frequency Machine Learning (RFML) introduces Machine Learning (ML) and Deep Learning (DL) techniques into modern radio communications systems, and is expected to be a core component of 6G technologies and beyond. While RFML provides a myriad of benefits over traditional radio communications systems, existing approaches are generally incapable of adapting to changes that will inevitably occur over time, which causes severe performance degradation. Transfer Learning (TL) offers a solution to the inflexibility of current RFML systems, through techniques for re-using and adapting existing models for new, but similar, problems. TL is an approach often used in image and language-based ML/DL systems, but has yet to be commonly used by RFML researchers. This dissertation aims to provide the foundational knowledge necessary to reliably use TL in RFML systems, from the definition and categorization of RF TL techniques to practical guidelines for when to use RF TL in real-world systems. The unique elements of RF TL not present in other modalities are exhaustively studied, and methods and metrics for measuring and predicting RF TL performance are examined.

Deep learning has become an ubiquitous part of research in all fields, including wireless communications. Researchers have shown the ability to leverage deep neural networks (DNNs) that operate on raw in-phase and quadrature samples, termed Radio Frequency Machine Learning (RFML), to synthesize new waveforms, control radio resources, as well as detect and classify signals. While there are numerous advantages to RFML, this thesis answers the question "is it secure?" DNNs have been shown, in other applications such as Computer Vision (CV), to be vulnerable to what are known as adversarial evasion attacks, which consist of corrupting an underlying example with a small, intelligently crafted, perturbation that causes a DNN to misclassify the example. This thesis develops the first threat model that encompasses the unique adversarial goals and capabilities that are present in RFML. Attacks that occur with direct digital access to the RFML classifier are differentiated from physical attacks that must propagate over-the-air (OTA) and are thus subject to impairments due to the wireless channel or inaccuracies in the signal detection stage. This thesis first finds that RFML systems are vulnerable to current adversarial evasion attacks using the well known Fast Gradient Sign Method originally developed for CV applications. However, these current adversarial evasion attacks do not account for the underlying communications and therefore the adversarial advantage is limited because the signal quickly becomes unintelligible. In order to envision new threats, this thesis goes on to develop a new adversarial evasion attack that takes into account the underlying communications and wireless channel models in order to create adversarial evasion attacks with more intelligible underlying communications that generalize to OTA attacks. / Master of Science / Deep learning is beginning to permeate many commercial products and is being included in prototypes for next generation wireless communications devices. This technology can provide huge breakthroughs in autonomy; however, it is not sufficient to study the effectiveness of deep learning in an idealized laboratory environment, the real world is often harsh and/or adversarial. Therefore, it is important to know how, and when, these deep learning enabled devices will fail in the presence of bad actors before they are deployed in high risk environments, such as battlefields or connected autonomous vehicle communications. This thesis studies a small subset of the security vulnerabilities of deep learning enabled wireless communications devices by attempting to evade deep learning enabled signal classification by an eavesdropper while maintaining effective wireless communications with a cooperative receiver. The primary goal of this thesis is to define the threats to, and identify the current vulnerabilities of, deep learning enabled signal classification systems, because a system can only be secured once its vulnerabilities are known.

While the toolset known as Machine Learning (ML) is not new, several of the tools available within the toolset have seen revitalization with improved hardware, and have been applied across several domains in the last two decades. Deep Neural Network (DNN) applications have contributed to significant research within Radio Frequency (RF) problems over the last decade, spurred by results in image and audio processing. Machine Learning (ML), and Deep Learning (DL) specifically, are driven by access to relevant data during the training phase of the application due to the learned feature sets that are derived from vast amounts of similar data. Despite this critical reliance on data, the literature provides insufficient answers on how to quantify the data training needs of an application in order to achieve a desired performance. This dissertation first aims to create a practical definition that bounds the problem space of Radio Frequency Machine Learning (RFML), which we take to mean the application of Machine Learning (ML) as close to the sampled baseband signal directly after digitization as is possible, while allowing for preprocessing when reasonably defined and justified. After constraining the problem to the Radio Frequency Machine Learning (RFML) domain space, an understanding of what kinds of Machine Learning (ML) have been applied as well as the techniques that have shown benefits will be reviewed from the literature. With the problem space defined and the trends in the literature examined, the next goal aims at providing a better understanding for the concept of data quality through quantification. This quantification helps explain how the quality of data: affects Machine Learning (ML) systems with regard to final performance, drives required data observation quantity within that space, and impacts can be generalized and contrasted. With the understanding of how data quality and quantity can affect the performance of a system in the Radio Frequency Machine Learning (RFML) space, an examination of the data generation techniques and realizations from conceptual through real-time hardware implementations are discussed. Consequently, the results of this dissertation provide a foundation for estimating the investment required to realize a performance goal within a Deep Learning (DL) framework as well as a rough order of magnitude for common goals within the Radio Frequency Machine Learning (RFML) problem space. / Doctor of Philosophy / Machine Learning (ML) is a powerful toolset capable of solving difficult problems across many domains. A fundamental part of this toolset is the representative data used to train a system. Unlike the domains of image or audio processing, for which datasets are constantly being developed thanks to usage agreements with entities such as Facebook, Google, and Amazon, the field of Machine Learning (ML) within the Radio Frequency (RF) domain, or Radio Frequency Machine Learning (RFML), does not have access to such crowdsourcing means of creating labeled datasets. Therefore data within the Radio Frequency Machine Learning (RFML) problem space must be intentionally cultivated to address the target problem. This dissertation explains the problem space of Radio Frequency Machine Learning (RFML) and then quantifies the effect of quality on data used during the training of Radio Frequency Machine Learning (RFML) systems. Taking this one step further, the work then goes on to provide a means of estimating data quantity needs to achieve high levels of performance based on the current Deep Learning (DL) approach to solve the problem, which in turn can be used as guidance to better refine the approach when the real-world data quantity requirements exceed practical acquisition levels. Finally, the problem of data generation is examined and provides context for the difficulties associated with procuring high quality data for problems in the Radio Frequency Machine Learning (RFML) space.

With the increase in spectrum congestion, intelligent spectrum sensing systems have become more important than ever before. In the field of Radio Frequency Machine Learning (RFML), techniques like deep neural networks and reinforcement learning have been used to develop more complex spectrum sensing systems that are not reliant on expert features. Architectures like Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) have shown great promise for applications like automated modulation classification, signal detection, and specific emitter ID. Research in these areas has primarily focused on "one size fits all" networks that assume a fixed signal length in both training and inference. However, since some signals are more complex than others, due to channel conditions, transmitter/receiver effects, etc., being able to dynamically utilize just enough of the received symbols to make a reliable decision allows for more efficient decision making in applications such as electronic warfare and dynamic spectrum sharing. Additionally, the operator may want to get to the quickest possible decision. Recurrent neural networks have been shown to outperform other architectures when processing temporally correlated data, such as from wireless communication signals. However, compared to other architectures, such as CNNs, RNNs can suffer from drastically longer training and evaluation times due to their inherent sample-by-sample data processing. While traditional usage of both of these architectures typically assumes a fixed observation interval during both training and testing, the sample-by-sample processing capabilities of recurrent neural networks opens the door for "decoupling" these intervals. This is invaluable in real-world applications due to the relaxation of the typical requirement of a fixed time duration of the signals of interest. This work illustrates the benefits and considerations needed when "decoupling" these observation intervals for spectrum sensing applications. In particular, this work shows that, intuitively, recurrent neural networks can be leveraged to process less data (i.e. shorter observation intervals) for simpler inputs (less complicated signal types or channel conditions). Less intuitively, this works shows that the "decoupling" is dependent on appropriate training to avoid bias and insure generalization. / Master of Science / With the increase in spectrum congestion, intelligent spectrum sensing systems have become more important than ever before. In the field of Radio Frequency Machine Learning (RFML), techniques like deep neural networks and reinforcement learning have been used to develop more complex spectrum sensing systems that are not reliant on expert features. Architectures like Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) have shown great promise for applications like automated modulation classification, signal detection, and specific emitter ID. Research in these areas has primarily focused on "one size fits all" networks that assume a fixed signal length in both training and inference. However, since some signals are more complex than others, due to channel conditions, transmitter/receiver effects, etc., being able to dynamically utilize just enough of the received symbols to make a reliable decision allows for more efficient decision making in applications such as electronic warfare and dynamic spectrum sharing. Additionally, the operator may want to get to the quickest possible decision. Recurrent neural networks have been shown to outperform other architectures when processing temporally correlated data, such as from wireless communication signals. However, compared to other architectures, such as CNNs, RNNs can suffer from drastically longer training and evaluation times due to their inherent sample-by-sample data processing. While traditional usage of both of these architectures typically assumes a fixed observation interval during both training and testing, the sample-by-sample processing capabilities of recurrent neural networks opens the door for "decoupling" these intervals. This is invaluable in real-world applications due to the relaxation of the typical requirement of a fixed time duration of the signals of interest. This work illustrates the benefits and considerations needed when "decoupling" these observation intervals for spectrum sensing applications. In particular, this work shows that, intuitively, recurrent neural networks can be leveraged to process less data (i.e. shorter observation intervals) for simpler inputs (less complicated signal types or channel conditions). Less intuitively, this works shows that the "decoupling" is dependent on appropriate training to avoid bias and insure generalization.

Recent innovations in machine learning have paved the way for new capabilities in the field of radio frequency (RF) communications. Machine learning techniques such as reinforcement learning and deep neural networks (DNN) can be leveraged to improve upon traditional wireless communications methods so that they no longer require expertly-defined features. Simultaneously, cybersecurity and electronic warfare are growing areas of focus and concern in an increasingly technology-driven world. Privacy and confidentiality of communication links are both more important and more difficult than ever in the current high threat environment. RF machine learning (RFML) systems contribute to this threat as they have been shown to be successful in gleaning information from intercepted signals, through the use of learning-enabled eavesdroppers. This thesis focuses on a method of defense against such communications threats termed an adversarial evasion attack in which intelligently crafted perturbations of the RF signal are used to fool a DNN-enabled classifier, therefore securing the communications channel. One often overlooked aspect of evasion attacks is the concept of maintaining intended use. In other words, while an adversarial signal, or more generally an adversarial example, should fool the DNN it is attacking, this should not come at the detriment to it's primary application. In RF communications, this manifests in the idea that the communications link must be successfully maintained with friendly receivers, even when executing an evasion attack against malicious receivers. This is a difficult scenario, made even more so by the nature of channel effects present in over-the-air (OTA) communications, as is assumed in this work. Previous work in this field has introduced a form of evasion attack for RFML systems called a communications aware attack that explicitly addresses the reliable communications aspect of the attack by training a separate DNN to craft adversarial signals; however, this work did not utilize the full RF processing chain and left residual indicators of the attack that could be leveraged for defensive capabilities. First, this thesis focuses on implementing forward error correction (FEC), an aspect present in most communications systems, in the training process of the attack. It is shown that introducing this into the training stage allows the communications aware attack to implicitly use the structure of the coding to create smarter and more efficient adversarial signals. Secondly, this thesis then addresses the fact that in previous work, the resulting adversarial signal exhibiting significant out-of-band frequency content, a limitation that can be used to render the attack ineffective if preprocessing at the attacked DNN is assumed. This thesis presents two novel approaches to solve this problem and eliminate the majority of side content in the attack. By doing so, the communications aware attack is more readily applicable to real-world scenarios. / Master of Science / Deep learning has started infiltrating many aspects of society from the military, to academia, to commercial vendors. Additionally, with the recent deployment of 5G technology, connectivity is more readily accessible than ever and an increasingly large number of systems will communicate with one another across the globe. However, cybersecurity and electronic warfare call into question the very notion of privacy and confidentiality of data and communication streams. Deep learning has further improved these intercepting capabilities. However, these deep learning systems have also been shown to be vulnerable to attack. This thesis exists at the nexus of these two problems, both machine learning and communication security. This work expands upon adversarial evasion attacks meant to help elude signal classification at a deep learning-enabled eavesdropper while still providing reliable communications to a friendly receiver. By doing so, this work both provides a new methodology that can be used to conceal communication information from unwanted parties while also highlighting the glaring vulnerabilities present in machine learning systems.