SMU awarded $4.2 million to advance AI-driven manufacturing and workforce training

SMU has received a $4.2 million award from the National Institute of Standards and Technology (NIST) to enhance digital twin technology and high-performance computing, strengthening the university’s role in developing the technologies and workforce needed to support the future of American manufacturing.

The two-year project will put a particular focus on semiconductor manufacturing, a strategically important industry for U.S. technological innovation and economic security.

Led by Nader Jalili, a mechanical engineering professor at the SMU Lyle School of Engineering, the project will support a new laboratory where students and researchers can explore how AI, robotics and humans can work together to improve manufacturing processes.

“If we want to bring more manufacturing back to the United States, we have to prepare the people who will make it work,” said Jalili. “That is where I see an important role for SMU’s research mission. This lab will allow us to work on manufacturing problems with industry while giving students and working professionals hands-on experience with the same technologies.”

The project will expand capabilities at SMU’s Center for Digital and Human-Augmented Manufacturing (CDHAM) to include a Digital Twin Technology Lab developed in collaboration with Mitsubishi Electric.

Paired with SMU’s high-performance computing, the lab will connect physical manufacturing with virtual models of these systems to better prepare SMU students and researchers for increasingly automated factories, where robotics and assembly processes are essential. They’ll explore how digital models, AI, robotics and automation can work together to improve manufacturing processes. 

“Human judgment and human connection cannot be replaced by AI,” Jalili said. “We are working on making the robot more responsive to the human presence.”

Jalili said the project reflects a shift toward manufacturing systems in which people and robots work together, rather than treating automation as a way to remove people from the process. By combining physical equipment with digital models, researchers can study how humans and robots can collaborate safely and efficiently.

Bridging the digital and physical worlds 

CDHAM specializes in digital twin technology, which creates a virtual representation of a physical system or component. With the help of SMU’s high-performance computing, including the NVIDIA DGX SuperPOD, these digital twins allow researchers to simulate operations, monitor performance and evaluate potential changes before implementing them in a real-world environment. The team aims to use this approach to improve operational efficiency by 10% to 20% while reducing downtime and waste.

The new Digital Twin Technology Lab will have a five-cell modular automation system featuring interconnected robotic workstations, conveyors and inspection stations that mirror processes used in semiconductor fabrication facilities and other advanced manufacturing environments.

This will allow students to follow a component through steps such as handling, processing and inspection, while studying how robotic systems and people interact along the way. Rather than working only with individual pieces of equipment, students will see how the systems function together. 

The lab will also support new digital twin courses, graduate research assistantships and industry-aligned training for the Dallas–Fort Worth semiconductor and advanced manufacturing workforce.

The lab’s digital twin will extend that experience beyond the physical line. Digital models can represent a larger factory environment, allowing students to explore systems and equipment that cannot all be housed in a university laboratory. Researchers can also use digital twins to examine manufacturing operations and test changes before applying them to physical equipment.

“This is a major change to what’s currently possible,” Jalili said.

Semiconductor assembly and automation without a cleanroom

The project takes a unique approach to semiconductor manufacturing education. The team will use aluminum wafers as lower-cost stand-ins for silicon wafers, allowing more students, researchers and trainees to work with manufacturing processes without the expense and specialized conditions of a semiconductor cleanroom. Researchers will also develop new ways to detect, diagnose and work around defects in chiplet-based advanced packaging to improve yield and reduce waste.

The project also will explore how robots can respond to people working nearby. Researchers aim to use information about a person’s attention and physiological state, along with task and environmental conditions, to help robotic systems adjust how they collaborate with a human worker. The broader objective is to improve safety and productivity while preserving the judgment and supervision people bring to complex tasks. In addition, the team will develop AI tools that turn plain-language instructions and camera images into robot action plans, checked against physical constraints and the digital twin before they run.

Jalili said this approach could help make manufacturing environments safer by enabling systems to anticipate potentially hazardous interactions and alert workers before an incident occurs.

Jalili, an ASME Fellow, focuses his research on intelligent manufacturing, AI-driven digital engineering, dynamics and control, and robotics. Helping Jalili create and run this two-year project are co-principal investigators Yildirim Hurmuzlu, Edmond Richer and Adel Alaeddini from SMU Mechanical Engineering and Jennifer Dworak from SMU Electrical and Computer Engineering. Their expertise spans control systems, robotics and automation, digital twin modeling, and semiconductor engineering.

This work was performed under the following financial assistance award 60NANB26D244 from U.S. Department of Commerce, National Institute of Standards and Technology.