Developing The Next Generation Of Army Tactical Vehicles

On the battlefield, soldiers must be able to trust their fellow teammates implicitly. Even if that teammate is a robot.

Pacific University alumna Maggie Wigness ’10 is at the forefront of making those robots trustworthy teammates for their human counterparts in the United States Army, bringing together artificial intelligence, computer vision and robotics to develop smart vehicles for military applications.

As a senior computer scientist at the U.S. Army Combat Capabilities Development Command (DEVCOM) Army Research Laboratory (ARL) robotic research group, Wigness and her team are charged with not only research but also fostering partnerships in the research community to broaden the Army’s access to expert talent and to accelerate transitions of science-enabled capabilities.

“I get to work with top academic researchers in the field,” Wigness said. “It’s enjoyable to engage with fellow researchers to collaboratively develop software and go out to the field together and test it. It’s fun to engage with the academic community and continue to push the research.”

Wigness’ research specialty is computer vision, where she uses artificial intelligence and machine learning techniques to allow computers or robots to see and understand the world around them. These algorithms convert the information about an environment to object classes, like trees or grass, that allows the robot to make intelligent navigation decisions, much like the human brain converts information sent to it through the eyes.

That information is important as Wigness’ group develops self-driving vehicles that can navigate in harsh conditions. Where commercially developed self-driving cars are navigating flat roads and traffic, these military vehicles must be able to navigate challenging terrain and rapidly changing environmental conditions.

“We have extreme elevation changes. We don’t have structured cues in the environment to guide us, such as lane markings or stop signs,” Wigness said. “How do we address the different terrains that are rough and rugged and difficult to traverse? How do we reason in these highly unstructured environments to provide information for these assets to move quickly and safely through an environment?

“It’s easy for people to understand self-driving cars and the autonomous car industry that the commercial sector is looking at. We’re trying to do something similar, but we’re doing it with smaller vehicles in much harsher conditions.”

While Wigness earned her PhD in computer science at Colorado State University in 2015, it was as an undergraduate computer science major at Pacific that she discovered her interest in machine learning and computer vision. The topics were not only discussed in her computer science classes at a time when both technologies were just starting to gain mainstream attention, but independent study and research courses allowed her to explore both topics in depth.

Those experiences led to a pair of projects that combined both topics with a passion for sports. For her senior project, Wigness investigated an algorithm that tracked the routes football players ran on the field, automating a critical part of the scouting process for players and coaches.

Wigness also conducted a summer undergraduate research project developing an algorithm to more accurately predict rankings of NCAA Division I-Bowl Championship Subdivision (BCS) football teams. The algorithm, which factored in specific plays in games as opposed to just location, outcome and strength of schedule, proved more accurate than the Bowl Championship Series, which was used at the time to determine the Division I-BCS champion.

Wigness believes that project, which earned her an invitation to present at the New England Symposium on Statistics in Sports at Harvard University, is what paved her way to graduate school.

“When you apply to graduate school, professors are looking for students who can do research,” Wigness said. “Because I had that experience, and because we published our algorithm at a sports statistics conference, schools had good evidence that I would be an asset to their research.”

As she has continued on her professional journey, Wigness has realized how much those two undergraduate projects are directly linked to the work she continues to do at the DEVCOM ARL. The subject matter is different, but the processes are quite similar.

“As my research career has progressed, I realize how important it is to be aware of the similarities across what may seem to be very different disciplines or applications,” she said. “As a researcher, I can be much more efficient with my contributions if I leverage the knowledge that my peers have already contributed to the research community. In a general sense, I can turn to the same underlying approaches that I used in my undergraduate efforts. It’s just the input and output that are changing.”

The challenge for Wigness and her fellow ARL researchers is not developing for now but for the future. Their research is designed to look 30 and 40 years down the road, trying to solve problems that most people are not thinking about now.

But whether now or in the future, the ultimate goal is to continue modernizing the military and ultimately saving lives.

“Ultimately, our research has a lot to do with keeping human lives further from the front line to keep them safe,” Wigness said. “So if we can throw automation, unmanned vehicles and unmanned robots into a situation before a human, then that’s a win.”