Dimitris Pados, Ph.D., principal investigator, professor, director of the CA-AI and a fellow of FAU I-SENSE in the Department of Electrical Engineering and Computer Science. (Photo by Alex Dolce)

In a landmark move to redefine artificial intelligence autonomous systems test and evaluation (T&E), Florida Atlantic University’s Center for Connected Autonomy and Artificial Intelligence (CA-AI) has secured a $799,759 grant from the United States Department of Defense (Air Force Office of Scientific Research). The grant will fuel the development of a cutting-edge platform for computational T&E of connected AI autonomy, solidifying FAU’s role as a national leader in next-gen networked AI autonomous systems research.

With this grant, FAU is poised to become one of the nation’s first research institutions to house a high-end NVIDIA GPU infrastructure for AI-driven autonomous system T&E.

Generative AI models – large language models such as GPT and Llama – are trained on enormous amounts of text and image data, largely gathered from the internet. These AIs have remarkable capabilities in producing human language and abstract concepts, but they’re limited in their grasp of the physical world and its rules.

Bringing AI autonomous systems from lab prototypes to real-world deployment is an intricate process riddled with safety and reliability challenges. Traditional physical testing is costly, time-intensive and constrained by logistical limitations, often restricting trials to a limited set of real-world scenarios. To build physical AI, teams need powerful, physics-based simulations that provide a safe, controlled environment for training autonomous machines. This not only enhances the efficiency and accuracy of robots in performing complex tasks, but also facilitates more natural interactions between humans and machines, improving understanding of spatial relationships and physical behavior of the 3D world that is inhabited.

FAU’s CA-AI will leverage the new funding to unlock new capabilities of physical AI by acquiring and developing state-of-the-art hardware and software infrastructure including:

• NVIDIA Omniverse infastructure to support the development of high-fidelity, physically based virtual environments required to represent the real environment and generate synthetic data necessary for training physical AI for robotics and next-generation wireless networks.
• Camera, LiDAR sensors, and AR/VR headsets to collect and render real video and image scenes, which could be then multiplied with 3D-to-real photo generation software in Omniverse.
• The NVIDIA DGX H200 platform to train or fine-tune AI models. Once trained, the model and its software stack can be validated in simulation using reference applications like NVIDIA Isaac Sim™ and NVIDIA Aerial Omniverse Digital Twin for 6G.
• NVIDIA Jetson platforms to deploy the optimized stack and policy model to run embedded in autonomous robots.

“This investment marks a significant step forward for AI innovation in Florida and beyond,” said Stella Batalama, Ph.D., dean, FAU College of Engineering and Computer Science. “By acquiring and developing state-of-the-art infrastructure, FAU’s Center for Connected Autonomy and Artificial Intelligence is establishing itself as a long-term leader in AI research. This technology will allow us to advance groundbreaking work in generative physical AI, test and evaluation of AI autonomous systems, fostering innovations that will have a lasting impact on both academia and industry at a national scale.”

Generative physical AI is unlocking new capabilities that will transform every industry. Building a high-fidelity virtual world environment will give FAU CA-AI the opportunity to test and evaluate the performance, reliability and validity of AI autonomous machines rigorously by subjecting them to a series of complex, dynamic scenarios that mirror real-world situations.

T&E challenges are exacerbated for connected AI autonomous systems involving extensive interaction with the physical world. The cyber-physical interaction via sensors and the system interactions with the environment are often difficult or impractical to create in the real-world for test purposes. To build and maintain the levels of trust and confidence required for the air and space forces of the U.S. Department of the Air Force to adopt AI autonomy technologies safely and responsibly, T&E must be built-in and ongoing. Examples of the needed T&E capabilities include instruments to record and transmit data that machine learning systems can easily consume, to detect deviations in the performance of AI-based systems for training and experimentation, synthetic data engines and digital twins, and the ability to regularly monitor and rapidly retrain and redeploy AI models. 

“AI has far-reaching implications for the air and space forces within the U.S. Department of the Air Force,” said Dimitris Pados, Ph.D., principal investigator, professor, director of the CA-AI and a fellow of the FAU Institute for Sensing and Embedded Network Systems Engineering (I-SENSE) in the Department of Electrical Engineering and Computer Science. “As AI technologies continue to evolve, the department will require dedicated test and evaluation capabilities, infrastructure and tools specifically designed for AI systems. These resources are essential to thoroughly assess the reliability, performance and practical utility of a wide range of AI-enabled systems. Ensuring these systems are rigorously tested will be critical for their integration into operations, where precision, trust and adaptability are paramount. By developing robust methods, policies and evaluation frameworks, we can help the Department of the Air Force harness the full potential of AI to strengthen and modernize national defense and space operations.”

Platform demonstrations will be made available to the public including organized visits from local high school students.

“Our platform will uniquely augment the capabilities of the team at the FAU Center for Connected Autonomy and Artificial Intelligence and enable new U.S. Department of Defense-related research opportunities such as test and evaluation of AI training datasets, learning with faulty and missing data, development of a digital spectrum twin for NextG communications and networking, simulation of schools of biorobotic fish or swarms of drones, among others,” said George Sklivanitis, Ph.D., co-principal investigator, Charles E. Schmidt Research Associate Professor and I-SENSE fellow, FAU College of Engineering and Computer Science.

CA-AI researchers will disseminate research results in major scientific conferences and journals and premier magazines in the field.

“The platform we develop will not only be accessible to colleagues and partners outside of FAU but also foster collaboration among principal investigators and faculty members within the College of Engineering and Computer Science and I-SENSE,” said Pados. “Furthermore, research outcomes will be integrated into both undergraduate and graduate curricula at FAU, including courses such as communication systems, engineering design I & II, information theory and smart antennas, ensuring that students gain hands-on experience with cutting-edge developments in the field.”

-FAU-