Interdisciplinary UAH faculty group wins $1.35M Army grant to advance human interactions with artificial intelligence

February 18, 2025 - UNTSVILLE, AL – A faculty group from The University of Alabama in Huntsville (UAH) has been awarded a $1.35 million grant from the U.S. Army Combat Capabilities Development Command Analysis Center to support advancements in human-systems integration analysis and artificial intelligence (AI). The interdisciplinary venture represents researchers from the departments of industrial and systems engineering, psychology and computer science at UAH, a part of The University of Alabama System, working together to improve human-AI interactions in AI-centric autonomous systems.

"This team has been working for many years on projects sponsored by the Navy, the Army, the Department of Defense and NASA," explains Dr. Bryan Mesmer, the principal investigator and associate professor of Industrial & Systems Engineering and Engineering Management (ISEEM). "Modern, complex systems, such as autonomous systems, are fundamentally interdisciplinary. To address the challenges associated with these interdisciplinary systems, it is necessary to approach them through an interdisciplinary lens."

In addition to Mesmer, UAH team members include Dr. Kristin Weger, associate professor of psychology; Dr. Lisa Vangsness, assistant professor of psychology; Dr. Vineetha Menon, an associate professor of computer science; and Dr. Howard Chen, Dr. Hanumanthrao Kannan and Dr. Ana Wooley, all assistant professors of ISEEM.

Human-systems integration (HSI) analysis is a systems engineering process that involves evaluating and integrating human factors – such as capabilities, limitations and performance needs – into the design and development of a system to ensure it is optimized for human use and minimizes potential errors by considering aspects like manpower, training, safety and occupational health throughout the system lifecycle.

"It is through the blending of the knowledge often trapped within these disciplines that we are able to make novel leaps in understanding to address the rapidly evolving needs of the Army and these complex systems they are adopting and are faced with," Mesmer notes. "It is through the strengths of the research team members and the support of the multiple colleges involved that we are able to effectively perform this interdisciplinary research."

Advancing human/AI collaborations

One of the specific topic areas to be examined by the grant involves a field called "human operator intent," which addresses the intended purpose or goal of a human when interacting with a system or performing a task. The topic is a key concept in fields like human-computer interaction and robotics where systems need to understand and anticipate user actions to provide effective assistance.

An area of emphasis along these lines is "human-AI teaming," where humans and AI combine their strengths in a synergistic fashion where humans and automated systems – like AI or robots – work together interdependently to accomplish shared objectives. The group will also explore "metacognition" – the processes used to plan, monitor and assess one's understanding and performance through an awareness of one's own thought processes and the patterns behind them. This approach is designed to determine whether human models can be applied to interactions with Autonomous/Artificial Intelligence Systems.

Another intriguing aspect of the effort is exploring the use of monocular 3D object detection and tracking, a kind of computer "vision" that can identify and locate objects in a 3D space using only a single 2D image from a camera by extracting depth information without additional sensors. This advance instead relies on visual cues like perspective, size and geometric relationships within the image.

"We are developing and validating a helmet-orientation tracking algorithm using a monocular camera and deep learning," Mesmer adds.

The group also aims to investigate implementing a digital twin-driven approach to AI model training. "Digital twins" provide virtual models of a physical object, system or process, using real-time data to simulate its behavior. Digital twins are used to monitor operations, identify potential faults and make better decisions in complex systems.

"We want to develop a framework for human intent inference and decision-modeling feedback for improved human-AI interactions in AI-centric assistive autonomous systems applications," Mesmer concludes. "The overall goal is to provide a rigorous, interdisciplinary foundation for the impact of the adoption of AI algorithms and autonomous systems."

This research is sponsored by the U.S. Army Combat Capabilities Development Command Analysis Center and is being accomplished under Cooperative Agreement Number W911NF-24-2-0158. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.