Project Description:

GTRI is building a simulation environment for mission autonomy that includes computer vision and digital scene generation. This project seeks to build a bridge between modern game engines (such as Unity and Unreal) with existing computer vision solutions (such an OpenCV and YOLO). The project will start with a proof-of-concept showcasing viability and expand to creating a vehicle tracking showcase in various environments.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Computer vision, simulation environment creation, software engineering

Project Description:

The ARCADIA project develops an innovative system where teams of robotic dogs work together to explore unknown environments while building interactive 3D virtual worlds. As the robots navigate through buildings or outdoor areas, they use advanced sensors and cameras to scan their surroundings and create a detailed digital replica. Users can put on VR headsets to virtually walk through these spaces in real-time, seeing what the robots are discovering as it happens, or explore the completed virtual environment later. This technology has important applications for reconnaissance, emergency response, and facility security, allowing people to understand dangerous or inaccessible spaces without being physically present. Students will work on cutting-edge robotics, artificial intelligence, 3D reconstruction, and virtual reality technologies while building a system that could save lives in real-world operations. The project offers hands-on experience with commercial robotic platforms, collaborative algorithms, and immersive visualization. By the end of summer, we will demonstrate multiple robots autonomously mapping a complex facility while users explore it in VR.

Citizenship Required:

None

Remote Work Allowed: No

Competencies:

Robotics, artificial intelligence, 3D reconstruction, virtual reality, immersive visualization

Project Description:

Hypersonic vehicles experience high surface temperatures during high-speed flight, which leads to a variety of complications. Research facilities housed within GTRI’s Cobb County Research Facility will be used (and new ones established) to study the damage that materials suffer when exposed to extremely high temperatures, and the effect of temperature on the structural dynamics of lightweight structures. This will include the use of a high-temperature plasma torch, optical assessments of thermal environments, and determination of material/structural responses through various means. Students will be expected to fabricate test articles, assemble new experimental setups and improve existing ones, and collect/analyze experimental data. Students will gain experience with operating a high-temperature inductively coupled plasma torch; conducting dynamic signal acquisition and processing; using high-speed cameras to obtain time-resolved optical flow visualizations; and studying vibrations of lightweight structures. Students will be exposed to advanced experimental methods and topics in materials science and structural dynamics, including material ablation and vibro-acoustics testing.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Materials science, structural dynamics, high-temperature testing, dynamic signal acquisition, optical flow visualization

Project Description:

This project aims to quantify the 'Utility' of a digital twin for training robot AI models, defined as the added performance gained by training the model with simulated vs. real data. Simulated data is much cheaper and more abundant to collect, but is limited in the extent to which it can model physical phenomena. The first core objective is to develop an understanding of how this utility varies depending on the specifics of an environment, task, and robot embodiment. The second objective is to develop strategies for optimal integration of simulated and real data (e.g. schedule, ratio, type) to jointly maximize performance and efficiency. Students will work with Nvidia IsaacSim to build and evaluate simulation environments, and will also compare to the performance of training entirely on physical robot hardware.

Citizenship Required:

US Persons Only

Remote Work Allowed: 100%

Remote Option: No

Competencies:

Artificial intelligence, robotics, simulation environments

Project Description:

We are exploring a novel method of bridging the gap between fully manual and fully autonomous operation using human-robot teaming and virtual reality systems. This allows us to rapidly deploy robots in new environments or performing operations where fully autonomous behavior is not currently possible. This not only gives us the ability to rapidly deploy robot systems, but also to bootstrap automation by crowd sourcing the data from the robot sensors and human decisions. Additionally, this framework allows for a one-to-many paradigm where a single operator can control teams of robots performing tasks. This project will explore methods to teach the robots to perform autonomously based on the data streams and human decisions. It also aims to demonstrate the one-to-many paradigm via a simulated application where a single user can control multiple systems in real-time.

Citizenship Required:

None

Remote Work Allowed: No

Competencies:

Human-robot teaming, virtual reality systems, autonomous robotics

Project Description:

The Water Plant Simulacrum will add an important industrial sector to the ICS Cybersecurity testbed, namely water treatment. Given the importance of clean water to all municipalities, as well as military bases, the addition of this utility to the testbed will provide GTRI with an excellent platform for developing and demonstrating novel cybersecurity methods and technologies. The control center will be implemented using Ignition, a commercial software package that is widely deployed in multiple industrial sectors globally. The physical simulation will be implemented using GRFICS-based applications provided by Fortiphyd, a local industrial cyber-security provider with origin at Georgia Tech.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Cybersecurity, physical simulation

Project Description:

This project investigates how to make Agentic AI systems — pipelines of multiple LLM-driven agents — more reliable and secure from two perspectives. First, we will formulate design guidelines that match Agentic workflows to the intrinsic limits of current LLMs, using information-theoretic analysis to bound the depth of sequential agent-to-agent conversations and context-window usage. Second, we will study how the compromise of a single agent (via prompt injection, hallucination, or other attacks) propagates through the system and degrades overall task performance. Based on these findings we will develop "zero-trust" guardrails that strictly validate inputs and outputs between agents, and quantify the resulting improvement in robustness. The deliverables are:

1. Best-practice recommendations for structuring Agentic pipelines
2. An empirical map of vulnerability propagation across agents
3. A plug-in framework of guardrails that can be added to existing workflows

The work will produce actionable insights for researchers and practitioners and give students hands-on experience with LLMs, Agentic AI, adversarial attacks, and advanced mathematics.

Citizenship Required:

None

Remote Work Allowed: Yes

Competencies:

Agentic AI, large language models (LLMs), cybersecurity

Project Description:

The Quantum Systems Division (QSD) performs advanced research on atomic systems, including trapped ion quantum computing and neutral atom platforms. This GRIP project aims to enhance the functionality, reliability, and integration of a critical hardware component, the Vescent qNimble digital PID controller, which is essential for controlling laser power — an indispensable aspect of QSD's research success. The qNimble is already a cornerstone of most QSD experiments, having demonstrated its capability to improve experimental outcomes since its adoption. However, the current in-house software and firmware supporting these controllers require modernization to eliminate runtime errors and better align with existing experiment control suites. As part of this project, a student-led effort will focus on refining firmware, developing seamless software integration, and constructing a small optical testbench to validate performance. Finally, comprehensive testing will be conducted on both neutral atom and trapped ion platforms, ensuring the qNimble operates efficiently within QSD's mission-critical research environments while reducing downtime and enhancing productivity.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Quantum computing, firmware development, software integration

Project Description:

This project will work to develop a Georgia Tech First Responder Ecosystem that is interoperable with city, state, and federal agencies. Georgia Tech is responsible for both student and the general public at sporting events and conferences and should have interoperable hardware and software to provide a Common Operational Picture (COP) to effectively coordinate and collaborate with supporting agencies on daily operations, (pre-planned) events, and (unplanned) incidents.

Citizenship Required:

None

Remote Work Allowed: No

Competencies:

software development, communications systems

Project Description:

This project establishes a database-backed digital turnover log to preserve technician tacit knowledge and maintain system annotations in a structured, quarriable format. It addresses a critical gap in current maintenance practice, where operational decisions are often undocumented, siloed, or buried in unstructured, handwritten formats—limiting traceability, disrupting continuity, and making it difficult to retrieve context-rich insights when they're most needed. The project targets TRL 6–7, with a validated prototype deployed in an operational environment and supported by documentation for transition. Deliverables include the database schema, input interface, and a speech to text capability.

Citizenship Required:

US Persons Only

Remote Work Allowed: No

Competencies:

Software development, databases

Project Description:

This project aims to develop micro-fabrication methods for single pixel and arrays of graphene IR light emitters arrays and investigate their radiance and response performances for IR scene projection applications. IR scene projectors (displays) are widely used in the Hardware-in-the-loop (HITL) testing of missile sensors. Current IR scene projectors are typically based on arrays of miniature resistive heater elements controlled in real time to display IR scenes to a system under test. Major disadvantages of these conventional thermal-emitter based scene projectors are the slow rise and fall times (~5 ms) due to the sluggish thermal response of resistive heaters, which results in limited frame refresh rate (~200 Hz). Also, thermal expansion of high-aspect-ratio resistor heater elements limit the maximum radiance/apparent temperature (<1000k) achievable by the resistor arrays. Graphene IR emitters can be developed with high temperature/radiance and fast thermal response, owing to the high thermal conductivity and ultrasmall mass of graphene array emitters.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

electrical engineering, mechanical engineering, materials science, physics

Project Description:

Modern FSOC systems emphasize an optical architecture to transmit information over long free-space distances through shift keyed signals. To understand performance degradation, many software-based simulators exist to predict atmospheric effects — namely fading due to turbulence, pointing, and jitter — to inform performance limitations yet are susceptible to model mismatch and fail to capture hardware nonlinearities. In this GRIP, we seek to develop a modular testbed built merging leading simulators with a hardware-in-the-loop framework to develop a research-grade tool to verify models, test innovations, and enable next generation functionality. This GRIP will comprise two student teams: the first team will integrate existing infrastructure throughout GTRI into a modular testbed and form the control backend to test a wide variety of experiments. The second team will utilize a leading wave optics turbulence simulator and arbitrary wavefront generator to deploy representative turbulence wavefronts for analysis. To push functional capabilities, both teams will interface with the emerging field of all-optical process (e.g. diffractive deep neural networks) by building on end-to-end architectures and integrating spatial light modulators to test learned beam shaping capabilities to create turbulence-robust optical information systems.

Citizenship Required:

US Persons Only

Remote Work Allowed: No

Competencies:

Optics, computer simulation, controls, physics

Project Description:

NOAA operates geostationary weather satellites that continuously detect lightning from space. However, the lighting sensors on these satellites are known to detect phenomena that are not lightning, such as explosions and satellites in lower orbits. There are also natural phenomena that resemble lightning but are not associated with severe thunderstorms, such as earthquake lights, which may be detectable from space. Finally, some lightning strikes as a "bolt from the blue," traveling through the clear air to strike targets far from thunderstorms. What all of these events have in common is that they are unpredictable and generally arise in locations without deep convective clouds. This project will make use of High Performance Computing (HPC) resources at GTRI to survey all instances of lightning-like events that are located in clear air or shallow cloud regions where electrification is not expected to occur.

Citizenship Required:

US Persons Only

Remote Work Allowed: No

Competencies:

Atmospheric sciences, high performance computing

Project Description:

Access to real in-field datasets for the purpose of research and analysis is limited by the expensive costs of hardware and field data experiments. This cost creates an entry barrier to experimental work for students and researchers outside the underwater acoustics community. While there are standard synthetic aperture sonar (SAS) dataset collections to work with, the underwater acoustic sensing community is limited in new and experimental data. The maritime sensors branch in the SEAL Lab has developed a Distributed In-air Ensonification and SEnsing Lab (DIESEL), a mature in air acoustic test bed that has demonstrated mastery in monostatic synthetic aperture sonar imaging. The motivation for developing this system is to explore opportunities to study bistatic and multistatic collections in an in-air, low-cost sonar test bed. Bistatic acoustic collections offer diversity in angle reflections beyond traditional monostatic acoustic reflections. Diversity in acoustic angle reflections tells us more about an objects shape and acoustic characteristics. Bistatic imaging coupled with monostatic imaging will expand the space of sensing information collected on underwater objects of interest and ordnance in mine counter measures to further the Navy’s objectives in successful detection and classification of objects on the seafloor. Students will work with the DIESEL testbed to simulate underwater sonar acoustic collections and apply signal processing and imaging processing to the DIESEL experimental datasets to further bistatic sensing and analysis.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Underwater acoustics, signals processing, image processing

Project Description:

This project will create a battle management command and control (BMC2) operational scenario within the Starcraft II (SC2) environment, that is centered on representing the decision making cycle (i.e. observe, orient, decide, act — OODA loop) of a mission crew commander (MCC) during dynamic re-planning in the course of critical decision points during mission execution. We will enable the SC2 scenario to be played in the MCC role, by a human player in the MCC role, a multi-agent player in the MCC role (i.e. an MCC digital twin — DT), or a human-agent team who both play the MCC role together. HSE is an established engineering process that describes the activities and artifacts supporting human-centered technology design for United States (U.S.) military applications, including but not limited to requirements generation, function allocation, and iterative HITL requirements-driven evaluations. The 2026 GRIP team will work in collaboration with an applied Master's program capstone team, to implement the HSE-driven scenario, MCC decision function allocation, and other relevant HSE specifications in the SC2 environment. In our current SEALC3D IRAD project, we are applying a human-centered approach to prototype a multi-agent model to assist the MCC of a four-person rotary wing aircrew assigned to C3 contingency operations. This proposed 2026 GRIP project will incorporate the final prototype of the MCC multi-agent model into the SC2 environment to learn and execute the BMC2 dynamic re-planning scenario, and ultimately partner in a limited HITL demonstration with the human player to jointly execute a selected set of the MCC decision functions specified in our BMC2 scenario.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Artificial intelligence, machine learning, data science, computer science, cognitive science, psychology, human-computer interaction

Project Description:

Many software packages have hundreds of dependencies; those dependencies can have hundreds of dependencies, and so on, forming a massive tree of open-source software. To better understand this tree of software and track potential supply chain risks, the Software Supply Chain Graph project seeks to assemble a graph database of information about the open-source software ecosystem. This project will heavily involve interfacing with online APIs using Python, the Neo4J graph database, and graph analytics algorithms. Students will learn about designing and maintaining ingests for OSINT (open-source intelligence), a skill that is broadly applicable both in the DoD and in the private sector. The project is based in Washington D.C., offering students a unique opportunity to meet government contacts.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Sotware engineering, computer science

Project Description:

Traditional spectrum monitoring systems often fail to provide real-time, actionable insights necessary to navigate a highly congested radio frequency environment, especially in scenarios involving sophisticated communication methods that involve variance in time, frequency, and space. These methods especially fail to allow for or near real-time decision-making in an electronic warfare environment that necessitates efficient decisiveness to maintain operational effectiveness. To address this gap, we propose Augmented Reality for Electromagnetic Spectrum Operations that integrates spatial array processing, higher order signal processing, and artificial intelligence for signal and channel characterization. To ensure efficacy of our designs, we plan to create a custom, reprogrammable radio frequency scene generator. Our generator will emit realistic electromagnetic scenarios that will rigorously train and test our system. This will help our models and agents avoid common issues affecting many artificial intelligence systems in use today, such as overconfidence and misleading perceptions. By leveraging state-of-the-art machine learning algorithms, multiple-input antenna sensing technologies and algorithms, and near real-time analytics, this system will deliver actionable and relevant information to enhance operational effectiveness at all levels of the command.

Citizenship Required:

US Citizens Only

Remote Work Allowed: No

Competencies:

Physics, electrical engineering, machine learning, computer science

Project Description:

Want to help bridge the skills gap for our nation’s veterans, military families, and first-generation college graduates through cutting-edge technology? Our team at GTRI is developing a GenAI-powered career readiness system designed to empower transitioning service members and their spouses as they navigate the evolving job market. Many job seekers with limited career readiness pathways, like veterans, struggle to translate their valuable experience into terms that resonate with civilian employers, while existing career tools often don't equip them to confidently use AI platforms that are increasingly relied upon in various industries. Our innovative solution focuses on implicitly building AI fluency—the ability to effectively collaborate with AI—by integrating it directly into the job search workflow, rather than through traditional training methods.

As a GRIP student, you’ll play a crucial role in developing and refining this system, tackling challenges in natural language processing, user interface design, and responsible AI development. This is a unique opportunity to contribute to a project that not only leverages advanced AI technologies but also genuinely impacts the lives of those who have served our country. Join us in this mission to address a critical problem facing our nation and help veterans and military families secure meaningful careers in a rapidly changing world while gaining hands-on experience with pioneering AI tools and methods.

Citizenship Required:

None

Remote Work Allowed: No

Competencies:

Machine learning, computer science, software engineering