Defense Notices

The pursuit of autonomy in cyber-physical systems (CPS) presents a challenging task of real-time interaction with the physical world, prompting extensive research in this domain. Recent advancements in artificial intelligence (AI), particularly the introduction of deep neural networks (DNNs), have significantly enhanced CPS autonomy, notably boosting perception capabilities. 

CPS perception aims to discern, classify, and track the objects of interest in the operational environment, a task considerably challenging for computers in three-dimensional (3D) space. For this task of detecting objects, leveraging lidar sensors and processing their readings with deep neural networks (DNN) has become popular due to their excellent performance. 

However, in systems like self-driving cars and drones, object detection must be both accurate and timely, posing a challenge due to the high computational demand of lidar object detection DNNs. Furthermore, lidar object detection DNNs lack the capability to dynamically reduce their execution time by compromising accuracy (i.e. anytime computing). This adaptability is crucial since deadline constraints can change based on the operational environment and the internal status of the system.  

Prior research aimed at anytime computing for object detection DNNs using camera images are not applicable when considered to lidar-based detection due to architectural differences. Addressing this challenge, this thesis focuses on proposing novel techniques, such as Anytime-Lidar and VALO (Versatile Anytime Lidar Object Detection). These innovations aim to enable lidar-based object detection DNNs to make effective tradeoffs between latency and accuracy. Finally, the thesis aims to integrate the proposed anytime object detection techniques into unmanned aerial vehicles and introduce a system-level scheduler capable of managing multiple anytime computation capable tasks.  

This thesis focuses on the multiple processing techniques that can be applied to a single receive element co-located with a Frequency Diverse Array (FDA) transmission structure that illuminates a large volume to estimate the scattering characteristics of objects within the illuminated space in the range, Doppler, and spatial dimensions. FDA transmissions consist of a number of evenly spaced transmitting elements all of which are radiating a linear frequency modulated (LFM) waveform. The elements are configured into a Uniform Linear Array (ULA) and the waveform of each element is separated by a frequency spacing across the elements where the time duration of the chirp is inversely proportional to an integer multiple of the frequency spacing between elements. The complex transmission structure created by this arrangement of multiple transmitting elements can be received and processed by a single receive element. Furthermore, multiple receive processing techniques, each with their own advantages and disadvantages, can be applied to the data received from the single receive element to estimate the range, velocity, and spatial direction of targets in the illuminated volume relative to the co-located transmit array and receive element. Three different receive processing techniques that can be applied to FDA transmissions are explored. Two of these techniques are novel to this thesis, including the spatial matched filter processing technique for FDA transmission structures, and stretch processing using virtual array processing for FDA transmissions. Additionally, this thesis introduces a new type of FDA transmission structure referred to as ”slow-time” FDA.

In response to the challenges posed by increasingly sophisticated criminal behaviour that strategically evades conventional identification methods, this research advocates for a paradigm shift in forensic practices. Departing from reliance on traditional biometric techniques such as DNA matching, eyewitness accounts, and fingerprint analysis, the study introduces a pioneering biometric approach centered on facial recognition systems. Addressing the limitations of established methods, the proposed methodology integrates two key components. Firstly, facial features are meticulously extracted using the Histogram of Oriented Gradients (HOG) methodology, providing a robust representation of individualized facial characteristics. Subsequently, a face recognition system is implemented, harnessing the power of the K-Nearest Neighbours machine learning classifier. This innovative dual-method approach aims to significantly enhance the accuracy and reliability of criminal identification, particularly in scenarios where conventional methods prove inadequate. By capitalizing on the inherent uniqueness of facial features, this research strives to introduce a formidable tool for forensic practitioners, offering a more effective means of addressing the evolving landscape of criminal tactics and safeguarding the integrity of justice systems. 

This paper introduces a novel approach to manage collections of artifacts through smart contract access control, rooted in on-chain role-based property-level access control. This smart contract facilitates the lifecycle of these artifacts including allowing for the creation,  modification, removal, and historical auditing of the artifacts through both direct and suggested actions. This method introduces a collection object designed to store role privileges concerning state object properties. User roles are defined within an on-chain entity that maps users' signed identities to roles across different collections, enabling a single user to assume varying roles in distinct collections. Unlike existing key-level endorsement mechanisms, this approach offers finer-grained privileges by defining them on a per-property basis, not at the key level. The outcome is a more flexible and fine-grained access control system seamlessly integrated into the smart contract itself, empowering administrators to manage access with precision and adaptability across diverse organizational contexts.  This has the added benefit of allowing for the auditing of not only the history of the artifacts, but also for the permissions granted to the users.  

Legacy radar systems largely rely on repeated emission of a linear frequency modulated (LFM) or chirp waveform to ascertain scattering information from an environment. The prevalence of these chirp waveforms largely stems from their simplicity to generate, process, and the general robustness they provide towards hardware effects. However, this traditional design philosophy often lacks the flexibility and dimensionality needed to address the dynamic “complexification” of the modern radio frequency (RF) environment or achieve current operational requirements where unprecedented degrees of sensitivity, maneuverability, and adaptability are necessary.

Over the last couple of decades analog-to-digital and digital-to-analog technologies have advanced exponentially, resulting in tremendous design degrees of freedom and arbitrary waveform generation (AWG) capabilities that enable sophisticated design of emissions to better suit operational requirements. However, radar systems typically require high powered amplifiers (HPA) to contend with the two-way propagation. Thus, transmitter-amenable waveforms are effectively constrained to be both spectrally contained and constant amplitude, resulting in a non-convex NP-hard design problem.

While determining the global optimal waveform can be intractable for even modest time-bandwidth products (TB), locally optimal transmitter-amenable solutions that are “good enough” are often readily available. However, traditional matched filtering may not satisfy operational requirements for these sub-optimal emissions. Using knowledge of the transmitter-receiver chain, a discrete linear model can be formed to express the relationship between observed measurements and the complex scattering of the environment. This structured representation then enables more sophisticated least-square and adaptive estimation techniques to better satisfy operational needs, improve estimate fidelity, and extend dynamic range.

However, radar dimensionality can be enormous and brute force implementations of these techniques may have unwieldy computational burden on even cutting-edge hardware. Additionally, a discrete linear representation is fundamentally an approximation of the dynamic continuous physical reality and model errors may induce bias, create false detections, and limit dynamic range. As such, these structure-based approaches must be both computationally efficient and robust to reality.

Here several generalized discrete radar receive models and structure-based estimation schemes are introduced. Modifications and alternative solutions are then proposed to improve estimate fidelity, reduce computational complexity, and provide further robustness to model uncertainty.

Austere environments are defined by the US Military as those regularly experiencing significant environmental hazards, have limited access to reliable electricity, or require prolonged use of body armor or chemical protection equipment.  We propose that in modern society, this definition can extend also to telecommunications infrastructure, areas where an active adversary controls the telecommunications infrastructure and works against the people such as protest areas in Iran, Russia, and China or areas experiencing conflict and war such as Eastern Ukraine.  People in these austere environments need basic text communications and the ability to share simple media like low resolution pictures.  This communication is complicated by the adversaries’ capabilities as a potential nation-state actor. To address this, Low Earth Orbit satellite clusters, like Starlink, can be used to exfiltrate communications completely independent of local infrastructure.  This, however, creates another issue as these satellite ground terminals are not inherently designed to support many users over a large area.  Traditional means of extending this connectivity create both power and security concerns.  We propose that Bluetooth Mesh can be used to extend connectivity and provide communications. 

Bluetooth Mesh provides a low signal footprint to reduce the risk of detection, blends into existent signals within the 2.4ghz spectrum, has security aspects in the specification, and devices can utilize small batteries maintaining a covert form factor.  To realize this security enhancements must be made to both the provisioning process of the Bluetooth Mesh network and a key management scheme that ensures the regular and secure changing of keys either in response to an adversary’s action or as a prevention of an adversary’s action must be implemented.  We propose a provisioning process using whitelists on both provisioner and device and uses attestation for passwords allowing devices to be provisioned on deployment to protect the at-risk population and prevent BlueMirror attacks.  We also propose, implement, and measure the impact of an automated key exchange that meets the Bluetooth Mesh 3 phase specification.  Our experimentation, in a field environment, shows that Bluetooth Mesh has the throughput, reliability and security to meet the requirements of at-risk populations in austere environments. 

Bulk Synchronous Parallel (BSP) applications comprise distributed tasks that synchronize at periodic intervals, known as supersteps. Efficient resource management is critical for the performance of BSP applications, especially when deployed on multi-tenant cloud platforms. This project evaluates and extends some existing resource management algorithms for BSP applications, while focusing on dynamic schedulers to mitigate stragglers under variable workloads. In particular, a Dynamic Window algorithm is implemented to compute resource configurations optimized over a customizable timeframe by considering workload variability. The algorithm applies a discount factor prioritizing improvements in earlier supersteps to account for increasing prediction errors in future supersteps. It represents a more flexible approach compared to the Static Window algorithm that recomputes the resource configuration after a fixed number of supersteps. A comparative evaluation of the Dynamic Window algorithm against existing techniques, including the Static Window algorithm, a Dynamic Model Predictive Control (MPC) algorithm, and a Reinforcement Learning (RL) based algorithm, is performed to quantify potential reductions in application duration resulting from enhanced superstep-level customization. Further evaluations also show the impacts of window size and checkpoint (reconfiguration) cost on these algorithms, gaining insights into their dynamics and performance trade-offs.

Degree: MS Project Defense (CS)

This ground-breaking project addresses a critical issue in crop production: precisely determining plant-available inorganic nitrogen (IN) in soil to optimize fertilization strategies. Current methodologies frequently struggle with the complexities of determining a soil's nitrogen content, resorting to approximations and labor-intensive soil testing procedures that can lead to the pitfalls of under or over-fertilization, endangering agricultural productivity. Recognizing the scarcity of historical inorganic nitrogen (IN) data, this solution employs a novel approach that employs Generative Adversarial Networks (GANs) to generate statistically similar inorganic nitrogen (IN) data. 

This synthetic data set works in tandem with data from the Decision Support System for Agrotechnology Transfer (DSSAT). To address the data's inherent time-series nature, we use the power of Long Short-Term Memory (LSTM) neural networks in our predictive model. The resulting model is a sophisticated and accurate tool that can provide reliable estimates without extensive soil testing. This not only ensures precision in nutrient management but is also a cost-effective and dependable solution for crop production optimization. 

Model Predictive Control (MPC) has emerged as a potent approach for controlling nonlinear systems in the robotics field and various other engineering domains. Its efficacy lies in its capacity to predictively optimize system behavior while accommodating state and input constraints. Although MPC typically relies on precise dynamic models to be effective, real-world dynamic systems often harbor uncertainties. Ignoring these uncertainties can lead to performance degradation or even failure in MPC.

Nonlinear latent force models, integrating latent uncertainties characterized as Gaussian processes, hold promise for effectively representing nonlinear uncertain systems. Specifically, these models incorporate the state-space representation of a Gaussian process into known nonlinear dynamics, providing the ability to simultaneously predict future states and uncertainties.

This thesis delves into the application of MPC to nonlinear latent force models, aiming to control nonlinear uncertain systems. We formulate a stochastic MPC problem and, to address the ensuing receding-horizon stochastic optimization problem, introduce a scenario-based approach for a deterministic approximation. The resulting scenario-based approach is assessed through simulation studies centered on the motion planning of an autonomous vehicle. The simulations demonstrate the controller's adeptness in managing constraints and consistently mitigating the effects of disturbances. This proposed approach holds promise for various robotics applications and beyond.

The advent of advanced natural language processing (NLP) techniques has revolutionized the way we handle textual data. This project presents the implementation of exploring contextual sentence understanding on the Quora Insincere Questions dataset using the pretrained BERT architecture. In this study, we explore the application of BERT, a bidirectional transformer model, for text classification tasks. The goal is to classify if a question contains hateful, disrespectful or toxic content. BERT represents the state-of-the-art in language representation models and has shown strong performance on various natural language processing tasks. In this project, the pretrained BERT base model is fine-tuned on a sample of the Quora dataset for next sentence prediction. Results show that with just 1% of the data (around 13,000 examples), the fine-tuned model achieves over 90% validation accuracy in identifying insincere questions after 4 epochs of training. This demonstrates the effectiveness of leveraging BERT for text classification tasks with minimal labeled data requirements. Being able to automatically detect toxic, hateful or disrespectful content is important to maintain healthy online discussions. However, the nuances of human language make this a challenging natural language processing problem. Insincere questions may contain offensive language, hate speech, or misinformation, making their identification crucial for maintaining a positive and safe online environment. In this project, we explore using the pretrained Bidirectional Encoder Representations from Transformers (BERT) model for next sentence prediction on the task of identifying insincere questions.