Dayton Engineer

By Shawn Robinson, associate director, UD news and communications

University of Dayton engineering researchers working with researchers at Wright-Patterson Air Force Base have patented a better way to track objects navigating through, around and between hidden spots. The system will improve security and search and rescue applications.

Video (.mp4) of the technology is here and here.

Similar to other systems, this automated system developed in part by UD Vision Lab Director Vijayan Asari and UD Research Scientist Theus Aspiras tracks an object moving through a scene where visibility is blocked by shadows, buildings and trees, among other things. Where their system improves on others is that it can remember the object's speed and trajectory to instantly re-track the object when the object re-emerges.

"If the system loses the object and doesn't have any idea where it has gone, the system automatically expands the area it's tracking and searches specific areas within that circle based on the object's previous speed and history, and predicted future path. It's like the system says, 'That vehicle should have come out here, but it could have come out there. It uses past performance as a predictor of future performance," Asari said. "The moment the object emerges from the shadow or whatever else was hiding it, the system will immediately reidentify the object and continue tracking it."

An example given by Asari and Aspiras included surveying a flood scene with rushing water. If a car or human goes under water or gets stuck in brush, the system automatically can alert the operator when and where that happened, and provide rescuers information on where that object possibly could come out based on speed and movement trends.

Even if two or three similar objects emerge, it's not a problem for the system to find the exact object it needs to find, according to Asari, because the system already has the object it's looking for in its memory. 

"It's exactly like what a human being does," Asari said. "It thinks ‘What do we expect to happen?’ If we don't see within an appropriate time frame where it was supposed to come out, our reaction is like, 'Where did it go?,' and you start looking everywhere you think it could be."

An improvement over previous systems, the system has the wherewithal to alert the human user there is a problem or failure. 

"Also like a human, the system can react, 'Uh, oh. I need to get help,'" Asari said. "Automated human perception is very much part of this technology."

In another improvement over existing technology, if the system does seek human assistance, it is designed to give as much information as possible to the user, so when they do come back, they are better oriented to the situation, according to Aspiras. 

"It's like if you're updating a co-worker about a problem, the system, almost like a briefing, will go through 'What happened? How fast were they going? What direction were they going?' The system not only notifies the user that it has failed but how the system has failed and provides diagnostic information," he said. "It almost functions like a black box in an airplane."

Asari and Aspiras have worked together since Aspiras' undergraduate days at Old Dominion University in the early 2000s. Aspiras followed Asari to UD's Vision Lab where he became one of many students whom Asari is proud to have mentored through the ranks. 

"Like Theus, several students start as undergraduates and use their experience in the lab as a catalyst to get a doctorate," Asari said. "We've even had a high school student do an internship, go elsewhere for his undergraduate degree, and then he came back for his master's degree and eventually a full-time job in Dayton while he worked on his Ph.D. Another student started in high school here as an intern and went to MIT."

But Aspiras never imagined when he started working with Asari that his name would be on a patent.

"I just knew I wanted to go as high as I could in an academic career, because I like to teach," he said. "I never thought I would go the full distance with Dr. Asari. It's been great. It's a long, tenuous process (to earn a doctorate), but it's incredibly rewarding."

For more information on the University of Dayton Vision Lab, visit https://udayton.edu/engineering/research/centers/vision_lab/index.php.