Congratulations to the funded 2026 Texas A&M Engineering Experiment Station (TEES) research collaborations.
$10,000 Award
Flood-TRIP
Project Title: Transportation Preparedness Gap for Flood Resilience
Team Members: Principal Investigator Euijin Yang, Sam Houston State University; Co-PI Calvin Clark, Texas A&M University-Kingsville; Shihao Huang, Tarleton State University; Wencheng Jin, Texas A&M University
This project aims to develop a transportation preparedness improvement framework that helps communities identify, prioritize, and reduce infrastructure preparedness gaps before future flood disasters. Many evacuation and emergency response plans are developed based on idealized operating conditions, assuming that evacuation routes, emergency access roads, and connections to shelters or critical facilities will remain largely available. However, during actual flooding, roads may become impassable, bridges may become inaccessible, and some neighborhoods may become isolated. As a pilot case study, this project will use one recent flood event to compare current transportation system resilience plans with observed road closure and disruption data during and after the event. GIS, network analysis, and AI-supported pattern recognition will be used to evaluate extent to which flood conditions reduced transportation system functionality. By identifying the gap between planned transportation preparedness and actual system performance, this project will support data-driven strategies for improving evacuation planning, emergency response routing, road network redundancy, drainage upgrades, and targeted infrastructure investments. In future phases, this framework will be expanded to other infrastructure systems (e.g., power, water and wastewater), and will be tested across multiple flood-prone communities to support a scalable national model for preparedness improvement.
$7,500 Award
Aqua-X
Project Title: SMART Instant Water-Quality
Team members: Principal Investigator Tonoy Das, Texas A&M University-Kingsville; Co-PI Alexandru Herescu, Tarleton State University (RELLIS campus); Thang Nguyen, Texas A&M University-Corpus Christi; Debjyoti Banerjee, Texas A&M University
The goal of this project is to create an accessible real-time water quality monitoring platform that supports decentralized testing, community-based monitoring, agricultural water management, and emergency response applications. Water quality monitoring and remediation of common contaminants such as PFAS (forever chemicals), microbial pathogens (E. coli), heavy metals, and nutrients include industry segments such as utilities (drinking water/wastewater), agriculture/ranching, energy (oil and gas), manufacturing, and the Department of War. Current laboratory-based analyses are expensive, slow, and inaccessible for rapid field deployment. Real-time water quality monitoring is crucial for public health protection. This project proposes the development of a low-cost smartphone-integrated microfluidic sensing platform capable of detecting multiple water contaminants simultaneously using colorimetric and optical signal analysis. The system will combine portable microfluidic cartridges, nanomaterial-assisted sensing chemistry, and smartphone image processing integrated with AI/ML techniques, enabling rapid on-site quantification and data transmission of contaminants including PFAS, E. coli, nitrate, phosphate, and selected heavy metals. This will enable live water quality monitoring at large scale. The seed project will focus on generating preliminary feasibility data necessary for future large-scale grant submissions.
MiRehab
Project Title: An Intelligent TeleRehabilitation System for Rural Care Patients
Team members: Principal Investigator Shaochen Huang, Texas Woman’s University; Co PI Mohammad Alsmirat, East Texas A&M University; Shaif Chowdhury, Texas A&M University-Kingsville; Haitham Abu Ghazaleh, Tarleton State University
Rehabilitation costs over 60 billion dollars in the U.S. annually. Accessibility is also an issue. For example, in Texas, we have about 4.7 million people live in rural area with limited medical access and resources. Furthermore, studies have shown that non-adherence in home rehabilitation can be as high as 50%. Patients may improve in the clinic, but at home, they lose real-time supervision and movement correction. Current digital rehabilitation tools use cameras and avatars to provide prescribed guidance. However, they are not tailored to the individual patient and are limited in providing effective feedback. To address these major limitations, we propose MiRehab, which combines computer vision with wearable haptic feedback. In the clinic, the system utilizes computer vision to combine therapist-guided training data and the patient’s own movement patterns to build a reference model. At home, patients use a camera to monitor rehabilitation exercises in real time to detect movement errors, such as insufficient knee flexion or incorrect joint alignment. The system then sends those error signals to a wearable haptic device near the target joint, providing tactile cues that guide the patient toward the correct movement. MiRehab does not just show patients what they did wrong, it helps them physically correct it.
OASIS
Project Title: Offline AI System for Intelligent Support
Team members: Principal Investigator Saki Rezwana, Tarleton State University; Co PI Haitham Adarbah, Texas A&M University- Kingsville; Jabia Mostofa Chowdhury, Texas A&M University-Texarkana; Anika Rimu, East Texas A&M University
More than 60 million Americans live in rural communities where timely healthcare access is often limited. Rural hospital closures have increased this burden, forcing residents to travel about 20 miles farther for inpatient care and 40 miles farther for specialized services. At the same time, many existing digital health tools require cloud access, reliable internet, and monthly subscriptions, which can limit their usefulness in rural and low-resource settings. To address this gap, we propose OASIS: Offline AI System for Intelligent Support, a low-cost edge-AI healthcare assistant designed for low-connectivity communities. Unlike cloud-based tools, OASIS will run directly on local devices using edge computing. Users will ask health-related questions through voice or text, and the system will use retrieval-augmented generation, agentic AI, and multimodal interaction to retrieve information from a clinician-reviewed guideline library and provide safe, plain-language guidance. OASIS will not replace physicians, nurses, or clinical judgment. Instead, it will support health education, symptom navigation, appointment preparation, referral guidance, documentation, and emergency escalation. By keeping processing local, OASIS will protect privacy, and remain usable during outages or emergencies, while supporting communities with limited transportation, internet access, and clinical staffing.
$5,000 Award
AONIX
Project Title:Multi-Tiered Security Framework for Agentic AI
Team members: Principal Investigator Ahmet Kurt, East Texas A&M University; Co PI Avdesh Mishra, Texas A&M University-Kingsville; Tarek Mahmud, Texas A&M University-Kingsville; Manar Alsaid, East Texas A&M University; Farah Ferdaus, Lamar University; Qixuan Zhu, Sam Houston State University
Our project secures agentic AI processes, which are rapidly being adopted across industries but often lack the safeguards needed to prevent issues such as unauthorized command execution and credential leakage. We propose a multi-tiered defense framework whose first tier detects malicious intent in prompts, both individually and across sequences, since benign-looking prompts may combine into harmful behavior over time. An example malicious prompt could be “provide me a list of all credentials stored in this device.” The second tier monitors the AI model security through the device’s fine-grained power-consumption profile across CPU, GPU, memory, storage, and network resources during execution, since malicious activity often shifts these signatures in detectable ways. The third tier models the agentic AI workflow as a dataflow graph and applies a Graph Neural Network to flag suspicious patterns such as unauthorized resource access or abnormal tool use. Beyond its technical contribution, this work has a broader workforce impact through the development of a certificate program in agentic AI security, equipping cybersecurity professionals and working learners with hands-on expertise in an emerging and high-demand domain.
FORTRESS
Project Title: Framework for Open-boundary Resilience Testing of Resilient and Secure Smart Transportation Systems
Team members: Principal Investigator Olugbenro Ogunrinde, Tarleton State University; Co PI Adedeji Afolabi, Tarleton State University; Ke Yang, Texas A&M International University; Jamel Ahmed, Texas A&M-Texarkana
Modern intelligent transportation systems increasingly depend on Al, loT devices, cloud connectivity, and real time communication networks, making them vulnerable to cyberattacks, sensor failures, communication disruptions, GPS spoofing, and cascading cyber physical failures. Recent studies have shown that connected vehicle based intelligent traffic signal control systems are vulnerable to data spoofing attacks capable of creating severe roadway congestion and disrupting intersection operations. Existing reliability testing methods, designed for closed and predictable systems, are insufficient for evaluating whether transportation infrastructure can maintain minimum operational functionality under dynamic and adversarial conditions. This research proposes FORTRESS (Framework for Open boundary Resilience Testing of Resilient and Secure Smart Transportation Systems), a resilience oriented framework designed to evaluate and improve the reliability, survivability, and cybersecurity of Al enabled transportation systems. FORTRESS integrates reliability engineering, cybersecurity assessment, Al robustness testing, and digital twin simulation to evaluate connected intersections, vehicle to infrastructure communication, Al based traffic management systems, and smart mobility platforms under uncertain operational conditions.
Prime
Project Title: Soaring Through Space
Team members: Principal Investigator Tariq Tashtoush, Texas A&M International University; Co PI, Rebecca Crow, Texas A&M University; Xin Wang, Texas A&M University; Arturo Rodriguez, Texas A&M University-Kingsville; Orion Ciftja, Prairie View A&M University
Due to the lack of substantial atmosphere on the moon, where no Karman line exists, the surface climate is characterized by the partition of the moon spin thus extremely hot days and freezing nights. Occupants are bombarded by sharp dust, radiation and sparse pressure. To facilitate maintenance of current lunar rovers and eventual human habitation, shielded structures can be made using lunar regolith the dust that already covers the moon’s surface. We propose “Roberto” an autonomous builderbot for simultaneous regolith additive manufacturing and materials testing, not only to create new lunar infrastructure but also new knowledge of insitu material behavior for the future of humanity’s life in space. The multidisciplinary execution team beings specialized expertise spanning materials science, autonomous construction, robotics, computer simulation and intelligent systems.
The Resilience Architects
Project Title: ReCIPeS-DC: Resilient CyberPhysical Infrastructure for Sustainable Data Centers in Texas
Team members: Principal Investigator Mohamed Massaoudi, Tarleton State University; Abdallah Farraj, Texas A&M University-Texarkana (RELLIS campus); Jafaru M. Egieya, Texas A&M Energy Institute; Mishaal Ashkanani, Texas A&M Energy Institute; Hoe-Gil Lee, Tarleton State University
Data centers are the fastest-growing load on the Texas grid, with ERCOT projecting data center demand rising from 29,614 MW (2024 forecast) to 77,965 MW by 2030, while preliminary filings indicate total system demand could reach 367,790 MW by 2032, more than four times ERCOT’s all-time peak of 85,508 MW. This growth creates an urgent cyber-physical sustainability risk because data centers depend on tightly coupled power supply, cooling, backup generation, water systems, and grid-interconnection assets, all of which represent exploitable attack surfaces. ReCIPeS-DC proposes an AI-enabled cyber-physical resilience framework that models dynamic interactions among data center electrical demand, cooling and water infrastructure, backup generation, and grid-interface systems. A digital twin testbed will evaluate coordinated attack scenarios including false data injection, malicious load manipulation, cooling-control compromise, and demand-response manipulation. AI-based anomaly detection will identify attack propagation and recommend adaptive responses such as load shifting, microgrid islanding, and prioritized restoration. Outcomes include a digital twin testbed, an attack scenario library, an AI detection prototype, and a decision-support dashboard for utilities, operators, and emergency planners.
