ÕýÆ·À¶µ¼º½

DALLAS (ÕýÆ·À¶µ¼º½) – Can AI and emerging technologies make airplanes safer?

A team of engineers at ÕýÆ·À¶µ¼º½’s Lyle School of Engineering is exploring how digital twins can bridge the gap between human expertise and digital intelligence during aircraft inspections – reducing errors, improving processes and ensuring consistency. The research was recently published in .

Digital twins are a virtual replica of a real-world object, system, or process that can evolve in real time through data. While not a brand-new concept, digital twins are quickly gaining traction and becoming more powerful though increased sophistication in AI. The human digital twin developed by ÕýÆ·À¶µ¼º½ researchers provides a digital representation of a human worker interacting with a manufacturing environment, then overlays digital intelligence to double-check quality assessments, scan for safety risks and gather data.

“Some studies in the aerospace manufacturing sector show that as much as 80 percent of all quality assessments are made subjectively by humans,” said Chris Colaw, ÕýÆ·À¶µ¼º½ faculty and Lockheed Martin Fellow. “Instead of replacing the human with automation, the human digital twin approach will enable industry to keep the human centered in the process and build their capability of performance through digital and technological teaming.”

ÕýÆ·À¶µ¼º½ researchers Garrett Madison, a Ph.D. student in mechanical engineering and math, Grayson Griser, a junior studying mechanical engineering and math, and Gage Truelson, a sophomore studying mechanical engineering, also contributed to Colaw’s digital twin research.

“The human digital twin augments the worker’s abilities and takes the human error out of it, especially for repetitive tasks,” Madison said. “If they’re tightening 100,000 bolts on fighter jets and forget to finish one, AI can detect the error and give them live feedback.”

Data-driven, digital approach to airplane inspections

Quality assurance engineers ensure that aircrafts meet strict standards for quality and safety – checking that bolts are tight, scanning for defects, looking for contaminants that might interrupt system functions, and more.

After each step in their checklist, they stamp an operations card to prove the task was completed. But there’s no proof that the step was completed at quality standards. By utilizing a human digital twin, workers can provide objective evidence that they’ve completed all steps in the process in compliance with safety and conformity requirements.

By connecting cloud-based security cameras, cloud-based virtual environments, virtual reality headsets and AI detection algorithms, their performance can be digitally documented and double-checked.

Here’s how the human digital twin works:

• The quality assurance engineer wears a virtual reality headset and views a virtual manufacturing environment. A digital checklist appears in front of them.
• As the engineer stamps the virtual checklist, AI detection algorithms confirm whether the step was completed correctly, delivering push notifications in real time if a safety issue is detected. The engineer can fix it on the spot.
• A cloud-based camera system monitors the workspace, documenting activity and providing timestamps for each step.
• Data streams provide detailed information on the duration of each step, process bottlenecks that need addressed, and more to improve efficiency.

“There is great business value in utilizing a digital approach that improves worker safety and provides recallable, objective evidence of each step in the process,” Colaw said. “This approach provides manufacturing companies with wisdom and insight to drive for performance improvements unlike what they can today.”

The digital twin is also a great tool for training on the job, Truelson said.

“By merging AR and VR, it’s as if someone is standing there telling you what you need to know in real time,” he said. “It can accelerate training many employees at once.”

Monitoring worker stress, strain and discomfort

Colaw says the human digital twin has potential for more than just identifying mistakes in quality assessments. By layering on additional components, such as biosensors, it can identify stress, ergonomic issues and emotional indicators that could affect performance or safety.

“Out of 20 steps on the checklist, you might determine that four of those steps elevate a worker’s heart rate. For example, we found that tightening bolts by hand creates high levels of strain,” Colaw said. “Knowing this we can now prioritize human well-being and potential, which positively impacts the manufacturing process.”

Colaw’s findings were presented at the 2025 AeroDef Manufacturing Conference, which brings together influential stakeholders advancing the aerospace and defense industries.