Funded programs 2023

1. Title: Managing Data and Resiliency for Autonomous Vehicles: Application in Search and Rescue
PI: Manish Bansal
Lead Institution: Virginia Tech
Funding Program: FY23 Seed Funding 

Summary:  In the past years, simultaneous occurrence of pandemic and hurricanes presented significant challenges to FEMA and therefore, FEMA is turning to robotic automation for variety of its activities. Though autonomous ground vehicle embedded with robotic cameras (AGV-RC) can augment the performance and safety of US&R task forces, the inadequate real-time data management are hindering the reliance on this technology. Specifically, a robotic cameras system requires huge amount of data processing and storage capacity to make real-time decisions for dynamic search operations, even though there is a possibility of redundant information provided by the cameras. The overarching objective of this research project is to develop novel mathematical optimization models and algorithms for solving stochastic combinatorial optimization problems that arise in planning teleoperations for AGV-RC.

2. Title: Novel Algebraic-Models for Resiliency Planning of Networked-Sensors and their Information Transmission Network from Cyberattacks
PI: Manish Bansal
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Research Engagement Program

Summary:
Networked sensors system enables multiple users and researchers to interact with a remote physical environment using shared resources. These sensors help in estimating environment states and also provide large streams of information/data to decision makers for conducting operations such as intelligence, surveillance and reconnaissance, in environment where it is tedious for humans to collect information. Significant hardware advances in sensing are enabling their reliable utilization. As a consequence, the need for sensors to save lives, time, and money has been increasing for a wide range of applications such as warehouse management, operations in mining and agriculture, disaster area surveying and mapping, search and rescue, and many more. These operations are maintained through effective information collection and propagation. This seed funding will furnish us with preliminary results and new interesting ideas which we plan to pursue over next three to five years.

3. Title: Distributed Space Adaptive Communications and Security for Multi-Constellation Networks
PI: Jonathan Black
Lead Institution: Virginia Tech
Co-PIs & Institution: Samantha Kenyon (VT)
Funding Program: FY23 Research 

Summary: Rapid growth in the rate of commercial space launches and operations are fundamentally changing the economics of Space and presenting new opportunities for addressing our most urgent societal needs. New satellite-enabled telecommunications (satcom) companies, government programs, and remote learning and working requirements create the framework for successful public-private partnerships (PPPs) led by universities to tackle hard problems such as unconnected/under-connected regions and communities and security in new networking paradigms. This project will collaboratively research new space-based high-bandwidth networks to address the cybersecurity challenges of inter- and intra-constellation communications of internet satellite constellations.

4. Title: Cybersecurity Threat within Southwest Virginia’s Agriculture
PI: Karen Carter
Lead Institution: University of Virginia at Wise
Co-PIs & Institution: David Frazier (UVA-Wise)
Funding Program: FY23 Research 

Summary: After years of traditional, generational farming methods, Southwest Virginia (SWVA) farmers are transitioning to smart technology utilization to assess and enhance value of their food production systems, in part as a possible avenue to meeting the global food demand. For the purpose of this study, smart technologies involve integrations of technology and data-driven agriculture applications to increase crop and herd production, along with increased quality. The proposed research is expected to gain insight in to SWVA agricultural producers’ awareness and utilization of smart technologies, as well as how to effectively determine training needs to assist in the access to services, personal information privacy (PIP), proprietary information safeguards, and IP address protection for SWVA producers.

5. Title: Trustworthy Services for Autonomous Mission Computing Systems
PI: Jin-Hee Cho
Lead Institution: Virginia Tech
Co-PIs & Institution: Bo Ji (VT), Thang Hoang (VT)
Funding Program: FY23 Research Engagement Program 

Summary: 
We will pursue a line of research for external proposals to build Autonomous Mission Computing Systems (AMCSs) that can ensure privacy-preserving and secure communications with trustworthy data in the presence of adversarial attacks. In addition, we will aim to provide ACMSs’ robustness against such attacks under high dynamics, various uncertainties, and severe resource constraints. Such AMCSs will be designed mainly to meet the needs of military or tactical teams assigned with time-sensitive, dynamic missions. Hence, we will target to submit our proposals to meet the needs of the Department of Defense or mission-critical or mission-oriented applications under severe constraints and sophisticated cyber and adversarial threats.

6. Title: 5G-BEACON: Blockchain Enhanced Architecture and Coding for Orchestrated Networks
PI: Maice Costa
Lead Institution: Virginia Tech
Co-PIs & Institution: Yalin Sagduyu (VT), Sachin Shetty (ODU)
Funding Program: FY23 Research 

Summary: 5G and beyond (nextG) systems will address the communication demands from numerous areas, such as entertainment, healthcare, manufacturing, agriculture, transportation, and public safety. In this project, we are particularly interested in their great potential to address the challenge of secure interoperability across networks for informed decisions. In the civilian domain, several applications can benefit from 5G for data-driven decision, including health monitoring systems, autonomous vehicles, and disaster response.

7. Title: Proposal team Building and Privacy Expert Recruitment for a Workshop on Bystander Obscuration in Wearable Augmented Reality Displays 
PI: Brendan David-John
Lead Institution: Virginia Tech
Co-PIs & Institution: Bo Ji (VT)
Funding Program: FY23 Research Engagement Program 

Summary:
The powerful suite of sensors on augmented reality (AR) devices is necessary for enabling immersive mixed-reality experiences. However, they are known to create unease in bystanders (i.e., those surrounding the device during its use) due to privacy concerns. A primary source of concern is the risk of identification and surveillance from recent advancements in facial recognition. Our solution is to use privacy-enhancing technologies (PETs) to protect bystander privacy while enabling AR's wide adoption and future. Our approach leverages a privacy-preserving API that processes raw data to remove identifying information about bystanders from the data before sharing sensor data to third-party applications, enabling AR apps while mitigating privacy risks for bystanders. The expected impact of this work is to ease the adoption of AR technologies, particularly in locations with strict privacy policies like GDPR, while enabling innovation for AR platforms and developers.  This project has three primary objectives: (i) identifying and building a team of experts necessary to holistically address bystander privacy, (ii) conducting a virtual workshop to engage expert researchers and practitioners with expertise in privacy across several disciplines while generating an understanding of settings in which PETs are a possible solution, and (iii) submitting an NSF proposal that addresses focused technical and societal challenges for large-scale deployment of AR PETs.

8. Title: Manufacturing Engineering Education Program Research Engagement
PI: Bradley Davis
Lead Institution: Virginia Tech
Co-PIs & Institution: Alan Michaels (VT)
Funding Program: FY23 Research Engagement Program

Summary:
The project will support travel expenses to develop contacts for a new Manufacturing Engineering Education Program built on an existing $3.8M grant from the Office of Naval Research as part of their Manufacturing Engineering Education Program (MEEP). The sponsors at DoD-OUSD have invited our team to submit for an expansion / renewal of the MEEP grants, this time with greater focus on secure wireless communications techniques and the broader impact of the DoD’s digital transformation. This program for Research Engagement program is for assistance in maximizing our probability of win for the submitted proposal by visiting the intended users of the MEEP technologies, who indirectly establish the demand functions for the OUSD decision makers.

9. Title: Medical Data Cybersecurity
PI:  Frank Della Pia
Lead Institution: GCAPS 
Co-PIs & Institution: none
Funding Program: FY23 Research Program 

Summary:
This is a research effort to produce a literature review centered on cybersecurity topics in healthcare, predominantly for protecting personally identifiable information, or PII. The information gathered throughout the literature review will serve both the wider medical research communities for better understanding methods for protecting PII and provide opportunities for GCAPS to consider methods of safely and effectively recording, storing, and disseminating medical data in future projects.

10. Title: Federated and Secure Spectrum Learning for NextG Communications Systems
PI: Tugba Erbek
Lead Institution: Virginia Tech
Co-PIs & Institution: Yalin Sagduyu (VT), Kai Zeng (GMU)
Funding Program: FY23 Seed Funding

Summary:  As Deep Learning (DL) becomes a core part of next generation (NextG) systems, there is an increasing concern about the vulnerability of DL to adversarial effects. In this context, smart adversaries may tamper with the training and/or test inputs to DL algorithms embedded in NextG communications. The problem of learning in the presence of adversaries is the subject to the study of adversarial machine learning (AML) that has received increasing attention in computer vision and NLP domains. The first research thrust in this project is to investigate the emerging attack surface of AML for Federated (FL) and corresponding attacks and defense schemes. Specifically, we will focus on the free‐riding attacks where some clients do not contribute to the FL model updates while still receiving the global model from the server. We will pursue developing a game‐theoretic framework to quantify the interactions among free‐riding and participating clients. We will start with establishing the non‐cooperative Nash equilibrium strategies and then extend the analysis to coalition games through cooperative game theory.

11. Title: Wireless Security with Data Oriented Modalities (WISDOM)
PI: Tugba Erbek
Lead Institution: Virginia Tech
Co-PIs & Institution: Yalin Sagduyu (VT)
Funding Program: FY23 Research Engagement Program 

Summary:
As we move towards a more connected world, wireless signal intelligence and communication applications have become increasingly important. While these applications benefit from the use of machine learning (ML), the complex decision space of ML make it susceptible to a new set of security challenges that need to be addressed before the safe adoption of the ML-driven wireless solutions by the industry. To address this need, we respond to Track 2: Startup engagement with the objective of generating novel means of research, innovation, or workforce development for wireless signal intelligence and security that will benefit both the academic and commercial ecosystems in Virginia. Our proposed research is to secure and protect ML-based wireless applications against adversarial exploits and attacks. Our vision is to leverage and advance our adversarial ML research solutions to transition them to the ML-based wireless product lines. This research engagement will build the foundations to detect vulnerabilities of ML-based wireless systems in both test and training times and protect them against a variety of attacks that target the ML engine in wireless applications.

12. Title: Distributed In-Network Authentication for Zero-Trust Security Enabled by Programmable Switches
PI: Peng Gao
Lead Institution: Virginia Tech
Co-PIs & Institution: Bo Ji (VT)
Funding Program: FY23 Research Engagement Program

Summary:
This project aims to design and develop a new software-defined perimeter architecture by leveraging programmable switches, to enable distributed, in-network authentication for enterprise and NextG networks. The PIs will collaborate with Digital Bazzar, a Blacksburg, Virginia-based software and Internet startup company that specializes in providing payment, identity, and credential technologies for individuals and businesses.

13. Title: Addressing the Workforce and Cybersecurity Challenges in Dairy Production
PI: Denis Gracanin
Lead Institution: Virginia Tech
Co-PIs & Institution: Gonzalo Ferreira (VT), Mohamed Azab (VMI), David Jones (VMI)
Funding Program: FY23 Research 

Summary: One of the major issues in modern dairy production and the agricultural industry is the lack of awareness of cybersecurity issues, and their impact on the industry and the workforce involved. The main challenges facing the industry are mainly the lack of awareness about such issues, their critical nature, and the fact that they can be used to induce severe harm. The goal is to study the potential scenarios and issues that can lead to cyber attacks or can be induced by a cyber attacker.  The investigation will lead the development of a set of training scenarios that can be integrated into the final product that helps elevate the worker’s awareness about cyber security, its impact, and how to detect and handle attack situations. Virginia Military Institute will lead the effort related to analyzing the dairy farm equipment infrastructure, and the related smart products being used in the dairy farm based on the previous research in cybersecurity and Internet of Things (IoT).

14. Title: Cybersecurity aspects of DA/NSF grants
PI: Denis Gracanin
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Research Program

Summary:
The adoption of healthcare technology is a complex process that requires major planning, implementation time, and continuous management (especially software updates). We witnessed in recent years a significant increase in healthcare related cyberattacks (e.g., ransomware attacks). Healthcare cybersecurity is essential for the normal functioning of healthcare organizations. There are many types of specialized healthcare information systems such as electronic health record (EHR) systems, e-prescribing systems, practice management support systems, clinical decision support systems, radiology information systems and computerized physician order entry systems. Cybersecurity breaches could result in identity theft, medical fraud, extortion, and the ability to illegally obtain controlled substances. The healthcare workers are also data-workers, who must interpret complex and dense amounts of data and seek meaning or patterns in large amounts of noise. They are increasingly dependent on the healthcare information systems and thus increasingly vulnerable to cybersecurity attacks. The introduction of new technological innovation has outstripped their understanding of the challenges for facilitating human-human relationships and cybersecurity factors. We propose to investigate the impact of cybersecurity factors on healthcare team workflows and determine the best practices in addressing the relevant cybersecurity problems. We will leverage our ongoing Destination Area 2.0 and National Science Foundation grants about future of the work in healthcare as a foundation for the following activities: expand our research contacts to include healthcare cybersecurity experts; conduct an in-depth systematic literature review; and write a report and identify the relevant federal funding opportunities.

15. Title: Secure and Self-powered Wireless Internet of Things (IoT) Devices Through Physical Unclonable Functions and Energy Harvesting
PI: Dong Ha
Lead Institution: Virginia Tech
Co-PIs & Institution: Sook Ha (VT), Fariborz Laharbi Pour (VT) 
Funding Program: FY23 Research Engagement Program

Summary:
Internet of Things (IoT) devices are pervasive these days to sense and exchange data, and pervasive IoT devices would be vulnerable to tampering. A physical unclonable function (PUF) is based on unique physical variations occurring naturally during semiconductor manufacturing and implemented usually in integrated circuits (ICs) and field programmable gate arrays (FPGAs). For a given input or condition, a PUF provides a "digital fingerprint" output. PUFs can be applied for high-security requirements including authentication and cryptographic key generation. The proposed research is to generate PUFs using energy harvesting transducers while powering IoT devices. The proposed research is well suited for secure and self- powered wireless IoT devices with applications such as bio-metric sensing for humans and farm animals. We will design three different secure wireless IoT devices based on the PUFs generated by energy harvesting transducers and prototype them using off-the-shelf components. The first device is powered by harvesting solar energy, the second one by vibration, and the last one by thermal gradient. We will test those devices in our lab and demonstrate the feasibility and practicality of the proposed research.

16. Title: New Cryptographic Audit Tools for Effective Data Integrity Attestation in Large-scale Storage-as-a-service Infrastructure
PI: Thang Hoang
Lead Institution: Virginia Tech
Funding Program: FY23 Seed Funding

Summary: The overarching objective of this proposal is to develop a series of new cryptographic data audit protocols for effective information security assurance, which not only addresses emerging soundness concerns of data outsourcing raised by the client but also achieves satisfactory efficiency (e.g., low computation cost, optimal audit tag size) for the service provider to comply with standard data regulations. The three key research thrusts are 1) Efficient cryptographic audit protocols for large-scale data integrity check, 2) Enable public auditability to improve transparency and further use-cases and 3) Formal security analysis, implementation, and experiments.

17. Title: Securing Time Synchronization and Applications to 5G/Next‐G
PI: Tom Hou
Lead Institution: Virginia Tech
Funding Program: FY23 Research

Summary: The objective of this project is to identify Precision Time Protocol's (PTP) vulnerability to insider adversaries in the IoT scenario and validate this vulnerability through lab‐based hardware experiments. Once the vulnerability is identified, we propose to develop a new byzantine‐resilient network time synchronization scheme, as an extension to the existing PTP. We will provide a rigorous proof on the correctness of our scheme and implement a proof‐of‐concept in our lab‐based testbed. More important, we will apply the new byzantine‐resilient network time synchronization scheme to our ongoing research on 5G URLLC in industry automation setting and 5G C‐RAN/ORAN.

18. Title: Supporting AI-Driven Clinical Applications: Novel Privacy-Preserving Training Methodology for Federated Learning on Sensitive Healthcare Data
PI: Ruoxi Jia
Lead Institution: Virginia Tech
Co-PIs & Institution: Derek Kaknes (VT)
Funding Program: FY23 Research Engagement Program

Summary:
In this project, we propose the development and testing of a privacy-preserving methodology for training large language models (LLM) on sensitive healthcare data.  This work builds upon the team members’ prior successes in implementing a modified version of the differential privacy (DP) system to preserve sensitive information in general language data.  If successful, this project will allow researchers to train LLMs and other AI-models on sensitive healthcare data, including electronic health records (EHR) data, without extracting sensitive information that could violate healthcare privacy regulations.  The capacity to train and retrieve a model on sensitive EHR data opens up immense opportunities for domain-specific model-tuning and federated learning across healthcare centers.  To capitalize on this technological advancement, this project also proposes the development of a collaborative incubator for AI-driven clinical applications comprising the teaching hospitals of Virginia Tech Carilion, University of Virginia (UVA) Health, and Virginia Commonwealth University (VCU) Health.  This will create a federated learning environment in which researchers can train, validate, and test data-driven applications across the EHR datasets of all three institutions; increasing generalizability and equity of novel AI-driven applications.

19. Title: Enhancing Cybersecurity of Power Systems with Reinforcement Learning
PI: Ming Jin
Lead Institution: Virginia Tech
Co-PIs & Institution: Peter Boling (VT)
Funding Program: FY23 Seed Funding

Summary: The project goals are to design, develop and validate algorithms for the detection of cyberattacks at Information and Communication Technology (ICT). By synergizing the optimization-based attack modeling with the inverse reinforcement learning (IRL) framework, we will develop a multi-stage detection method that continuously monitor ICT network data to capture complex sequential anomalies with interpretable severity metrics to inform proper mitigation measures.

20. Title: Security and Privacy Aware Testbed for Voice-based Social Networks
PI: Youna Jung
Lead Institution: Virginia Military Institute
Co-PIs & Institution: Mohamed Azab (VMI), Denis Gracanin (VT)
Funding Program: FY23 Research

Summary: The goals of this project are to explore the voice-based social networks (VBSNs) topics 1) Blockchain-based decentralized identity and reputation management for VBSNs, 2) Cloud-based voice recognition, voice modulation, and automatic translation between text and voice, 3) Open-source testbed of privacy-preserving VBSNs and to produce a testbed that allows researchers and developers to learn security and privacy issues of VBSNs and in turn evaluate and test their research outcomes or products. This project will significantly advance social networking systems and cybersecurity.

21. Title: Quantum Computations for Smart Electric Grids: Enhancing Situational Awareness and Securing Power Systems Operations
PI: Vasileios Kekatos
Lead Institution: Virginia Tech
Co-PIs & Institution: Jamie Sikora (VT)
Funding Program: FY23 Seed Funding

Summary: We propose to develop, analyze, and evaluate quantum and quantum-classical algo­rithms for enhancing cyberphysical security and efficiency of electric power systems. The novel algorithms aim to: i) expedite monitoring functionalities (power system state estimation (PSSE), bad data processing, anomaly detection); and ii) scale up optimal power flow (OPF), a challenging yet omnipresent optimization task, to power networks having thousands of nodes.

22. Title: Enhancements of SQISign
PI: Jason LeGrow
Lead Institution: Virginia Tech
Co-PIs & Institution: Travis Morrison (VT)
Funding Program: FY23 Research Engagement Program

Summary:
SQISign is a recently-proposed post-quantum cryptographic protocol. We will explore the mathematics of SQISign in order to improve efficiency and/or security of SQISign, or to introduce advanced functionalities that can enhance user privacy when using SQISign.

23. Title: Resurrecting SIKE: Developing and Implementing New  Isogeny-Based-Quantum Schemes”  
PI: Jason LeGrow
Lead Institution: Virginia Tech
Co-PIs & Institution: Krzysztof Gaj (GMU)
Funding Program: FY23 Research Program

Summary: After the recent break of supersingular isogeny key establishment (SIKE)—once the most promising isogeny-based protocol—some attack countermeasures have been proposed. These new countermeasures create dramatic inefficiencies in the protocol. We propose to study techniques for optimizing isogeny evaluation, develop new algorithms, and provide state-of-the-art implementations of new isogeny-based key establishment protocols to determine their viability.

24. Title: Learning the Attackers’ Behavior for Defense of Smart Power Infrastructures
PI: Chen-Ching Liu
Lead Institution: Virginia Tech
Funding Program: FY23 Research

Summary: The electric power grid plays a pivotal role in modern life. From industrial work to domestic comfort, people depend on the power grid to provide electricity for varying purposes. In the wake of climate change and the accompanying quest to electrify transportation and heating, the electric power grid must take on an even more crucial role. Consequently, the cyber security, resilience, and controllability of the power grid is of paramount importance. The quest for a more reliable and resilient grid has led to the deployment of the smart grid, which is essentially a power grid enhanced with information and communications technology (ICT) to allow for remote control of the wide area grid. While this has made the smart grid a preferrable choice over the traditional grid, it has also made the grid vulnerable to cyberattacks that could lead to cascading failures, equipment damage, widespread and long-lasting outages, and adverse impacts on human lives and the economy. In this project, we will develop a novel decentralized technique for learning the attacker’s behavior, and hence, their motive in a cyberattack.

25. Title: Secure, distributed computing and communication
PI: Gretchen Matthews
Lead Institution: Virginia Tech
Funding Program: FY23 Research

Summary: This project centers on the use of coding theory to support security in two domains: distributed matrix multiplication and post-quantum cryptography. While the topics may seem disparate at first glance, there is commonality in the tools that we employ to address them. The first objective of this project is to design algorithms which allow a central node to provide a user the product of two matrices A and B which is obtained using N servers in such a way that no information about the matrices A and B is revealed even if T servers collude. The second objective of this project is to determine efficient decoding algorithms for the multivariate Goppa codes, which would provide an alternative to classic McEliece with a smaller public key size.

26. Title: Quantum Keys Management for Satellite Communications
PI: James McClure
Lead Institution: Virginia Tech
Co-PIs & Institution: Samantha Parry Kenyon (VT), Denis Gracanin (VT), Mohamed Azab (VMI), David Jones (VMI)
Funding Program: FY23 Research Program

Summary: The objective of this project is to develop modeling capabilities for quantum key distribution (QKD) in satellite networks and to perform vulnerability assessments related to their impact on cybersecurity. In contrast with classical encyrption techniques, which are known to be vulnerable to future quantum computers, quantum keys rely on fundamental laws of physics to provably detect eavesdropping. In terrestrial networks QKD can be performed using fiber-optics, which have a limited range due to signal losses. Free-space optical links in satellite networks are able to overcome these limitations, which make them an important developmental area for future communication systems.  

27. Title: Graduate Research & Innovation Design: Secure Connectivity for Renewable Devices in the grid (GRID-SECURED)
PI: Ali Mehrizi-Sani
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Research Program

Summary:
The electric energy delivery system, commonly known as the power grid, is a critical infrastructure and a component of the national security.  Transition to renewable energy resources, which can be deployed in a distributed manner close to power consumption sites, has necessitated the availability of communication infrastructure. Such infrastructure additionally has the potential to enhance the performance of control and protection facilities in the power grid. However, this also makes the power grid an attractive target for cyber-attacks. The overarching objective of this project is to create and demonstrate the GRID-SECURED framework. GRID-SECURED is a scalable, distributed, and cost-effective technology to enable secure and resilient communication to control and monitor distributed energy resources (DER), including grid-forming (GFM) inverters, within the power system and/or microgrids.

28. Title: InVerters for cleAN HOlistic Electricity Security (PREP-IVANHOE)
PI: Ali Mehrizi-Sani
Lead Institution: Virginia Tech
Co-PIs & Institution: Chen-Ching Liu (VT), Agnieska Miedlar (VT)
Funding Program: FY23 Research Engagement Program 

Summary:
The overarching goal of this research effort is to create a holistic design with theoretical guarantees for cybersecurity and cybersecure controls of an inverter-integrated power system. Our design will start with creating multi-timescale-enabled models that will be used for cybersecure controllers for the inverters. Considering the large number of entities that need to be coordinated and controlled together, the control of assets typically becomes a distributed control problem that needs to ensure cybersecurity. Our work will make foundational contributions to guarantee the performance requirements of such cybsersecure controls. Our specific objectives are 1) Create physically inspired data-driven multiscale models to represent the power grid; 2) Design security-embedded controls for inverters; 3) Create algorithms to mitigate cyberintrusions in the absence of well-structured data; and 4) Train the next generation of diverse cyberaware U.S. workforce.

29. Title: Microgrids for Autonomous Delivery of Secure Energy (MADS-ENERGY)
PI: Ali Mehrizi-Sani
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Research Program 

Summary:
Electric energy is a major enabler of the national economy: about one-third of the energy consumed in the U.S. industrial, commercial, and residential sectors is in the form of electrical energy. This has made the power grid an attractive target for cyberattacks, Efforts are underway to enhance the cybersecurity of the national power grid. A simultaneous and parallel (not competing) solution that can help achieve the security and resiliency needed by the grid is the concept of microgrids. A microgrid is a small-scale power system that has well-defined geographic boundaries and can operate both connected to and independently of the main grid. Microgrids are of particular interest to the Commonwealth of Virginia because of the military bases, data centers, and federal establishments in NoVA that can be operated more securely and efficiently as microgrids. This project proposes a two-pronged approach: (i) distributed cyberthreat detection and mitigation: We will create both topology-aware (physics-based) and topology-agnostic (machine learning-based) tools to detect and mitigate a broad range of cyber intrusions., and (ii) secure distributed control and power-sharing: We will build on our constant-frequency operation paradigm based on angle droop and implement the cybersecurity applications to include a layer of defense at the converter level to detect/mitigate sensor errors that can arise from a compromised firmware update, communication spoofing, e.g., false data injection, or synchronization attacks.

30. Title: SmartGuide: Revolutionizing the Depth and Dependability of Vision-Impaired Navigation
PI:  Na Meng
Lead Institution: Virginia Tech
Co-PIs & Institution: Luis Borunda Monsivais (VT), Andrew Gipe-Lazarou (VT)
Funding Program: FY23 Research Engagement Program 

Summary:
Globally, at least 2.2 billion people live with vision impairment, amounting to an annual economic impact of $411 billion in lost productivity. Although the majority of people with vision impairment and blindness are over the age of 50, vision loss continues to meaningfully affect people of all ages. Traditional navigation methods like guide dogs, canes, and GPS-based navigation tools provide some assistance to visually impaired people for exploring previously unknown areas/spaces. But these methods are limited in three major aspects. First, they are incapable of generating complex interactions which could facilitate and enrich users’ navigation experience. Second, they provide limited functionalities of sensing or reasoning about the surrounding environment, and thus cannot always vigilantly identify hazards or prevent upcoming dangers. Third, for emerging mobile technologies in particular, they lack accessible security mechanisms to protect users’ data or privacy. Consequently, users may spend excessive time and effort trying to unlock their apps, or else altogether deactivate built-in security mechanisms to make apps more accessible. We will create a new audio navigation system, SmartGuide, which will smartly interact with visually impaired individuals to help them explore unknown spaces in a safer, securer, and more meaningful way.

31. Title: Interactive story circle workshops to increase communities’ cybersecurity preparedness and prevent scams targeting vulnerable populations
PI: Katalin Parti
Lead Institution: Virginia Tech
Co-PIs & Institution: Susanna Rinehart (VT)
Funding Program: FY23 Research Program

Summary:
This project seeks to organize interactive story circle workshops to increase communities’ cybersecurity preparedness and prevent scams targeting vulnerable populations. This current project utilizes a theatre-devising process known as the story circle: a practice originating out of numerous indigenous traditions and developed by theatre practitioners to create performance and community “telling and listening” projects (von Rotz & Tokarski, 2020; Peterson, 2017). Specifically, we will organize and facilitate story circles (workshops) at Virginia Tech among and within international students, employees and community members; and individuals with immigrant backgrounds to come together to share experiences with cyberscams. The goal of the workshops (story circles) is to provide safe space for these groups to share their scam-related experiences and possible victimizations to scams. The research will reveal specific scam scenarios international and immigrant individuals experience, and shed light on their needs, and resources available to cope with scams. Through story circles (workshops), it will (1) raise awareness about specific scams victimizing immigrant/international people, (2) empower said individuals by creating a community of people with similar experiences (scam-victimization), and (3) help VT develop resources and better serve said individuals. 

32. Title: Secure & Privacy-Preserving Deep-Learning Framework for Smart and Connected Communities
PI: Angelos Stavrou
Lead Institution: Virginia Tech
Co-PIs & Institution: Xinghua Gao (VT), Thinh Doan (VT)
Funding Program: FY23 Seed Funding 

Summary: Smart and sensor-enabled buildings have emerged as the next generation of smart interfaces for humans to interact with the environment, technologies, and each other. People are spending 87% of their time inside buildings for work, rest, and entertainment. Currently, building technologies are evolving towards the integration of physical, digital, and human systems in the building environment to deliver a sustainable, prosperous, and inclusive future for people. There are, however, some major challenges in this domain primarily pertaining to the security, privacy, and use of the human and building generated information. Our interdisciplinary research team attempts to address these challenges by introducing a secure and privacy preserving data collection and sharing platform aimed at indoor sensing data for smart and connected residential communities. Using this platform, we aim to foster the development of novel deep-learning data analysis approaches that would balance multiple sustainability requirements for modern residential communities leveraging building sensors and actuators.

33. Title: 5G and 6G Security Through Deception
PI: Nishith Tripathi
Lead Institution: Virginia Tech
Co-PIs & Institution: Jeff Reed (VT)
Funding Program: FY23 Research

Summary: This project proposes a way to deal with cyber risks through deception rather than traditional encryption. Deception in the network can help protect a user's data and find and analyze an adversary. It can be applied at various points within the network. Deception has been used in Wi-Fi networks, an example being a "honey pot," which appears to be a vulnerable access point set up to monitor an intruder. A honey pot is a way to distract an intruder and learn more about the intruder. This project proposes to apply this defensive strategy to the 5G/NextG network.

34. Title: Securing the Interactions with AI-based Question-Answering Dialog Systems
PI: Bimal Viswanath
Lead Institution: Virginia Tech
Co-PIs & Institution: Megan Duncan (VT)
Funding Program: FY23 Research Program 

Summary:
Recent advances in Large Language Models (LLMs) have enabled high-quality question-answering (QA) and dialog systems. As these are largely trained on human-human conversation data, any imperfections in the training dataset can lead to biased or harmful responses by these systems. This project aims to systematically characterize and mitigate vulnerabilities in such LLM-based interactive systems. We focus on two key aspects of these systems---toxicity and truthfulness. This project has three main research thrusts: (1) Our aim is to systematically measure toxicity and truthfulness in publicly available dialog systems. (2) We are developing ML-guided approaches to mitigate toxicity and untruthfulness in dialog systems. Our key idea here is to leverage LLMs as few-shot learners to reason about toxicity and truthfulness. (3) Finally, we will develop new datasets to detoxify and improve truthfulness in dialog systems.

35. Title: Trustworthy Multimodal Machine Learning in Healthcare: Aligning Model Attention with Human Attention
PI: Xuan Wang
Lead Institution: Virginia Tech
Co-PIs & Institution: Aiguo Han
Funding Program: FY23 Research Engagement Program

Summary:
This research focuses on enhancing the security and trustworthiness of machine learning models used in healthcare. These models, known as multimodal ML models, combine different types of data such as images, text, and video to improve their performance. While these models hold great potential for transforming healthcare, they also present challenges related to security and trustworthiness. One major concern is the lack of interpretability in these models, making it difficult to understand how they arrive at their decisions. This lack of transparency can be problematic, especially if the model makes errors that could harm patients. To address this issue, the researchers propose aligning the model's attention with human attention during the training process. Currently, researchers analyze the model's attention, which highlights the important parts of the input data for decision-making. They have observed similarities between the model's attention and human attention, providing a basis for interpretation. However, few studies have attempted to align the model's attention with human attention during training. We hypothesize that aligning model attention with human attention during training will significantly improve the performance, interpretability, and trustworthiness of multimodal ML models in healthcare. The short-term objective of this research is to develop a reliable framework for achieving this alignment. This framework will contribute to the creation of more reliable and secure models, ultimately benefiting both patients and healthcare providers.

36. Title: High Accuracy Automatic Code Repair for Mission-critical Software
PI: Danfeng (Daphne) Yao
Lead Institution: Virginia Tech
Co-PIs & Institution: Bimal Viswanath (VT), Ismini Lourentzou (VT)
Funding Program: FY23 Research 

Summary:  Studies have shown that writing cryptographic code is error-prone, even for experts. Despite recent advances, state-of-the-art automatic code completion solutions have multiple deficiencies. Most data-driven code embedding solutions are not designed for addressing security-sensitive code. It is unclear how effective these approaches are in cryptographic code. In addition, there has not been any systematic investigation of various designs or comprehensive evaluation in terms of their security and accuracy capabilities. In this proposed project, our team will explore code embedding approaches and novel neural network designs for high accuracy code completion. Our solutions will assist software engineers with the development of security-critical codebases.

37. Title: Towards Large-Scale Human-AI Interactive Curriculum Generation for Online Learning
PI: Dawei Zhou
Lead Institution: Virginia Tech
Co-PIs & Institution: Shuo Yao (RU)
Funding Program: FY23 Research Program 

Summary:
Our proposal directly addresses the interest of CCI Call (securing the interactions between humans and machines) by proposing a human-AI interactive curriculum generation framework for labor-optimized and trustworthy online learning platforms. We tested and refined our prototype in PI's online classes in 2023. During the past few months, my student, Shuaicheng Zhang, has published on paper in ICML 2023.

38. Title: Towards Open Knowledge Network Construction, Adaptation, and Deployment
PI: Dawei Zhou
Lead Institution: Virginia Tech
Co-PIs & Institution: Junjie Hu (WI)
Funding Program: FY23 Research Engagement Program 

Summary:
This project aims to develop a unified data infrastructure for open knowledge network. The PI is currently working on a proposal to NSF IIS programs.

1. Title: Enhancing Cryptography Education Using Collaborative Visual Programming: A workforce development approach
PI: Sherif Abdelhamid
Lead Institution: Virginia Military Institute
Funding Program: FY23 Workforce

Summary: Cryptography is the science of securing sensitive information and ensuring that only the intended recipients can access and process the encrypted data. Internet shopping, online payments, and social networking websites have become increasingly popular with the advancement of the internet. However, hackers are getting more skilled than before to exploit existing vulnerabilities and attack these websites. Due to this, it has become increasingly important to introduce the science of cryptography to future generations, at a younger age, in straightforward and more engaging ways.  As a response, we are implementing a web-based programming learning tool called vizLab. The tool will help students bridge the gap between cryptography's mathematical foundations and computing by using a visual approach to programming. Students will learn to construct data encryption algorithms with minimal programming experience, using graphical icons representing the language’s essential elements. vizLab can execute the students’ block-based algorithms online; also, it can translate them into a high-level programming language (Python). Additionally, vizLab can store the completed blocks within an online cloud database. Students can share the constructed blocks with peers working on the same projects. Finally, students can integrate vizLab with learning management systems (LMS) to share their work with their instructors for assessment.

2. Title: Pathways for Cyberbiosecurity Workforce Preparation: Integrating Insights from Both Cybersecurity and Biosecurity
PI: Eric Kaufman
Lead Institution: Virginia Tech
Co-PIs & Institution: Feras Batarseh (VT), Anne Brown (VT), Susan Duncan (VT), Heather Lindberg (VWCC), B Bagby (VWCC)
Funding Program: FY23 Workforce

Summary: Cyberbiosecurity is an emerging field at the interface of the life sciences and the digital world and workforce development in cyberbiosecurity is a critical need in Southwest Virginia. Our specific aims for this project are: (1) crosswalk educational standards for biosecurity and cybersecurity into a comprehensive and integrative framework for cyberbiosecurity education, and (2) synthesize/analyze stakeholder perceptions that may guide curricular planning for cyberbiosecurity education.

3. Title: Broadening Participation in Security
PI: Gretchen Matthews
Lead Institution: Virginia Tech
Funding Program: FY23 Workforce

Summary: This project will broaden the participation in cybersecurity education, research, workforce development, and innovation. It will connect people in related areas to enhance opportunities and build capacity, working to (1) identify target groups, (2) determine metrics for success, (3) create points of entry for engagement and participation, (4) generate increased participation in events and programming, and (5) strengthen relationships with stakeholders to broaden impact and reach. We aim to increase diversity in terms of gender, race, ethnicity, geographic origin throughout Southwest Virginia, socioeconomic background, levels of learning (middle school, high school, community college, 4-year instituitions), disciplines, thus expanding opportunities for research funding, furthering curriculum development, diversifying the node student body and ultimately the cybersecurity workforce.

4. Title: Women in Security
PI: Gretchen Matthews
Lead Institution: Virginia Tech
Funding Program: FY23 Workforce

Summary: Women in Security is based on the idea that mentoring has been identified as one of the key contributors to the success of women in STEM disciplines. A number of activities will take place through the year to build and strengthen the mentor relationships. The activities are designed to build community and create connections amongst those involved. They will expose the undergraduates to what research and graduate school are like in a supportive environment where mentors are encouraged to openly and honestly share their experiences. They will help students, undergraduates and graduate students, to see themselves in a cybersecurity career through increasing the swath of role models available to them. They will enhance the relationships of postdocs and early career faculty, broadening the scope of what they see as possible and equipping them with knowledge shared by those more senior in the field.

5. Title: Cyber VIP: Use & Abuse of Personal Information
PI: Alan Michaels
Lead Institution: Virginia Tech
Funding Program: FY23 Workforce

Summary: The Use & Abuse of Personal Information (U&A) experiential learning effort engages a diverse multi-disciplinary group of undergraduate students to explore and quantify how personal information propagates across the Internet. The proposed CCI effort builds upon 2 years of experimentation that demonstrated the ability to generate realistic fake identities, perform one-time online interactions, and subsequently collect and analyze how that information is being both used and abused across email, SMS text, and voicemail modalities. Of particular interest are cross-site sharing behaviors (attributable due to one-time interaction), adherence to published privacy policies, trends across industries, root sources of spam / malicious content, and answering a variety of social science questions.

6. Title: Virginia Cybersecurity Education Conference Teacher Sponsorship
PI: David Raymond
Lead Institution: Virginia Tech
Funding Program: FY23 Workforce

Summary: The Virginia Cyber Range has built a strong user community in Virginia public high schools and colleges and we serve thousands of users in hundreds of Virginia schools each semester. Each year we host the Virginia Cybersecurity Education Conference, giving high school and college educators in the Commonwealth an opportunity to share ideas and continue to build the Virginia cybersecurity education ecosystem. This year, we are working with CCI partners in the Coastal node to host the conference on the campus of Old Dominion University. Many high school cybersecurity educators do not have travel funds available. In an effort to defray costs for high school teachers whose schools are unable to pay for their conference attendance, we would like to partner with CCI to fund their participation.

7. Title: HBCU Quantum Partnership Workshop
PI: Wayne Scales
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Workforce Program 

Summary:
The 2023 workshop will build on successful funding proposals that were submitted as a result of the 2022 workshop. These proposals teamed Virginia Tech and Virginia State University (VSU) to establish a QISE experiential learning laboratory at VSU in 2023 and also ongoing funding proposals with both VSU and Prairie View A&M University (PVAMU) to establish partnerships in QISE enabled technologies in cybersecurity, communication systems, artificial intelligence, machine learning, computing, and sensing. These partnerships are being designed to expand the research capacity and workforce development at HBCUs in QISE and QISE enabled technologies which is a priority of the federal government and industry. In fact, the 2022 and 2023 workshops will be a strong selling point for the ongoing funding proposals. Best practices for equitable HBCU partnering will result from the 2023 workshop since more careful input data from a larger group of HBCUs will be collected. Speakers from Virginia Tech, HBCUs, and other leading QISE research organizations will participate in the 2023 workshop.

8. Title: Resilient and Secure BattleDrones: Drone Racing League for Southwestern Virginia
PI: Kevin Schroeder
Lead Institution: Virginia Tech
Co-PIs & Institution: Jonathan Black (VT), Mohamed Gebril (GMU)
Funding Program: FY23 Workforce 

Summary: The work proposed here seeks to expand the CCI BattleDrones Competition beyond the inaugural year, expanding the number of institutions participating in the competition, offering the returning teams the opportunity to  incorporate lessons learned from the first competition while integrating new cyber-based obstacles and testing cyber resiliency techniques. The proposed project will therefore work with multiple members of the SWVA node, and will be scalable to other institutions around the Commonwealth. The second year of the competition will incorporate new challenges as each team will be responsible for designing the drone autonomy with the appropriate resiliency to endure various cyber obstacles such as intermittent GPS, spectrum denial, loss of camera visuals, signal spoofing, etc. Multiple successive races in a league format allows for developed drone technology, algorithms, and other existing cyber and cyber-physical tools to complement an autonomous drone system, demonstrating cyber effects and resiliency on public and private drone systems. Students will be encouraged to understand the existing infrastructure and to find ways and means to increase their chances of surviving and outperforming their opponents.

9. Title: Undergraduate Research Experience
PI: Prem Uppuluri
Lead Institution: Radford University
Funding Program: FY23 Workforce

Summary: This project will allow Radford University undergraduates to work on a CCI SWVA research projects at Virginia Tech. The students will be participating for eight hours a week for the fall semester of 2022 and the spring semester of 2023. They will be fully integrated into the project. Team members will be researching, proposing, and refining research questions answerable using the Open Source Intelligent (OSINT) collection engine. Questions may span any legal academic discipline, but will be evaluated on expected interest, feasibility of answering, and scientific approach by faculty experts.

1. Title: Protecting Bystander Visual Data Privacy in Augmented Reality Systems
PI: Bo Ji
Lead Institution: Virginia Tech
Co-PIs & Institution: Brendan David-John (VT)
Funding Program: FY23 Innovation Program

Summary:
In this project, we aim to design, develop, and prototype the first practical bystander privacy protection system that can effectively protect bystander visual (camera and depth) data privacy in real-time without negatively impacting user experience in Augmented Reality (AR) systems. The product will be a privacy-preserving API that sanitizes bystander information from sensor data streams before they are accessed by third-party applications. At a high level, it modifies how third-party applications access raw visual data, identifies and obscures the bystander’s faces, and passes on the obscured frames to the application.

2. Title: Power system reLiAnt Technology for Massive INterconnection of Integrated ReNEwables (PLATFORM-9)
PI: Ali Mehrizi-Sani
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Innovation Program 

Summary:
The transition to renewable energies is of course exciting, but it presents a humongous challenge for the power grid operations (utilities, independent system operators [ISO], and regulatory bodies). Before any new generation device (e.g., solar farm) is allowed to connect to the grid, its interactions with the grid under several scenarios must be thoroughly studied. These engineering studies are called “generation interconnection studies" (GIS) as defined by IEEE/ANSI standards and FERC/NERC (Federal Energy Regulatory Commission) regulations. The studies seem to be straightforward, but they are not. First, they are repetitive and time-consuming; some simulations take hours to complete and there are hundreds of such simulations, each varying one or two parameters, that are needed to complete the study. Second, data and system models are often not complete or are out of date, which leads to an inconsistent and incoherent representation of the system; finding these inconsistencies is difficult. Third, the US grid has system operators that are geographically dispersed; while they all operate under the FERC mandates, their operational practices sometimes differ significantly. Keeping track of such industry knowledge is a mobility constraint. Our goal is to create an automated system for (i) validating the consistency of the models and data, and (ii) performing the mandated studies (e.g., GIS) and additional advisory studies (e.g., cybersecurity studies). As of now, no company has a ready-to-use simulation platform and in-house expertise for all the electricity markets in the U.S. Similarly, no company exists that is specialized in running these studies and can provide a holistic and A-Z solution for all the electricity markets. At the same time, it is too early and risky for the big companies to fully engage in developing such a simulation platform.

3. Title: Anti-Counterfeiting Authentication App Using Deep-Learning-Based Physically Unclonable Functions
PI: Emma Meno
Lead Institution: Virginia Tech
Co-PIs & Institution: Mark Thompson (VT) 
Funding Program: FY23 Innovation Program

Summary:
Counterfeiting is an increasingly prevalent global issue, threatening the health and safety of individuals, corporations, and nations. Several legal measures have been implemented to secure product legitimacy but fall short of offering a user-end safeguarding mechanism. This project proposes a mobile application to authenticate products using previously researched security labels. These labels utilize an innately random property of silk proteins to create a unique identifier called a physically unclonable function (PUF). A machine-learning-based key generation algorithm, involving some image pre-processing steps, deep-learning prediction, and randomness extraction techniques, is then used to generate a key of 0s and 1s. This binary key is suitable for cryptographic applications. The objectives of this project are to determine a sufficiently secure, suitable authentication protocol for validating generated keys and build a reliable, resilient user-end mobile application prototype for validating PUF label authenticity. In the previous phase of this project, the cryptographic keys generated from the unique PUFs were sufficiently random and thus cryptographically secure, evaluated using the NIST SP 800-22 Statistical Test Suite. The impact of the proposed project is driven by human factors and global security, aiming to safeguard industries and consumers from counterfeiting threats. The expected outcomes from this work are to survey bio-PUF authentication protocols and to prototype a mobile authentication app using both uploaded and phone-camera-captured PUF label images.

4. Title: Cyber RADaR (Cybersecurity Rapid Asymmetric Discovery and Reporting): Intellectual Property (IP) Registration
PI: Jeff Pittges
Lead Institution: Radford University
Co-PIs & Institution: Bobby Keener (Civilian Cyber)
Funding Program: FY23 Innovation Program

Summary:
The first phase of the Cyber RADaR project produced a proof of concept and the second phase will develop a prototype with input and guidance from industry partners. The Cyber RADaR: IP Registration project will provide initial protection for the intellectual property (IP) used by the CyberRADaR projects. As the project team begins to consult with external partners and begin product development, intellectual property must be protected.  The project will protect IP by preparing and submitting trademark and provisional patent applications with the United States Patent and Trademark Office (USPTO). Trademarks and patents will be filed by CivilianCyber, which will offer IP protection for two brands: (1) CyberRADaR, and (2) the underlying WIaaS technology. CivilianCyber will work with Radford University to resolve any IP concerns. The project team has IP registration capabilities on staff and has added an IP registration expert consultant (Michael Miller) for further support. Each step, for both trademarks and patents, requires USPTO approval to continue the process and the process will end either when USPTO fully approves or denies the request. The primary performer of each step is listed after the step name but note that the project team will be in constant communication with the USPTO throughout the process to provide any feedback/clarifications.

5. Title: Cyber RADaR (Cybersecurity Rapid Asymmetric Discovery and Reporting) Phase Two: Commercially Deployed Beta Product
PI: Jeff Pittges
Lead Institution: Radford University
Co-PIs & Institution: Bobby Keener (Civilian Cyber)
Funding Program: FY23 Innovation Program

Summary:
Most cyberattacks are inherently asymmetric, targeting security vulnerabilities that “inflict a proportionally large amount of damage compared to the resources needed” to commit the attack. Until a patch is available, interim remediation to combat asymmetric cyber threats can be particularly complex for professionals, especially in small and medium sized organizations. Technology and cybersecurity professionals have increasingly turned to online social media to share and consume the latest threat information with a particular focus on understanding and remediating zero-day threats. Using online social networks is not new. In fact, social networks are becoming the primary source for threat remediation support. To better understand and address the need to use social network data for cyberthreat identification and remediation, our project team completed an initial project sponsored by SW CCI: Cyber RADaR: Cybersecurity Rapid Asymmetric Discovery and Reporting via AI-driven Social Media Crowdsourcing. During the phase one project, the team developed and validated a Cyber RADaR proof-of-concept (PoC) using advanced data processing techniques, including machine learning (ML) and natural language processing (NLP), to demonstrate the capability to automate the discovery, collection, transformation, analysis, and presentation of online social network (twitter and Reddit) sourced asymmetric threat data. The Cyber RADaR Phase 2 project seeks to enhance and expand phase one capabilities. Efforts to accomplish this will include: 1) additional development of new and existing algorithms; 2) functional and user experience (UX/UI) dashboard enhancements; and 3) the establishment of a customer feedback group to iteratively test and recommend product modifications.

6. Title: Scalable Continuous Monitoring Solutions for Enterprise Security
PI: Daphne Yao
Lead Institution: Virginia Tech
Co-PIs & Institution: none
Funding Program: FY23 Innovation Program

Summary:
This commercialization effort aims to develop deployable continuous monitoring solutions for managing enterprise security against stealthy external and internal threats. The specific product that we envision is a piece of stand-alone monitoring software, tentatively named SoftAuditor. SoftAuditor aims to detect stealthy attacks and anomalies in an organization, e.g., due to advanced persistent threats (APT) such as SolarWinds APT, as well as malicious insiders that may cause data leak. SoftAuditor can be deployed on endpoints or in the cloud, making intelligent security decisions based on observed employee and system data. Our ongoing efforts are focused on developing solutions to support better forensic capabilities that are urgently needed for cybersecurity investigation by security analysts. Our work leverages our existing malware detection expertise, as well as findings from our prior market research. Regarding potential clients and possible distribution mechanism, one direction is to explore the possibility of integrating our solutions to existing IT security frameworks and products as an add-on cybersecurity extension.