Students aim to defuse deadly traffic-stop interactions

Two events that happened within two days of each other got the attention of chemistry major Vaibhav Tadepalli ’17 and chemistry Ph.D. student Chris Reyes. But they had nothing to do with chemistry.

On July 6, 2016, during a routine traffic stop, a Minnesota police officer shot and killed Philando Castile; on July 8, Missouri police officer Michael Flamion was shot and paralyzed as he walked toward his car during another traffic stop.

“That got us thinking,” Tadepalli recalls. “There’s got to be a better way to do this. There’s got to be a way to put some sort of separation to keep nerves and bad situations like this from happening.” The problem was that fraught first moment: A police officer had to look into a stopped car, which might contain a weapon; the person in the stopped car had to wonder whether a nervous police officer might misperceive a situation or react according to fear, prejudice, or poor training. “This initial first contact,” Tadepalli says, “is the most dangerous.” There needed to be a way for the officer and the motorist to have their first interaction, after which everybody could calm down. “And I just said, ‘What if we could build a robot?’ ”

A few weeks of 3D printing wheels and body parts later, the two had a proof-of-concept chassis: the Sentinel, a little fourwheel robot, about the size of a microwave, with an accordion- style extension lift. It drives to the driver’s door and raises a computer tablet up to seven feet, high enough to reach the biggest SUV window. Tadepalli and Reyes (and their fledgling company, Trapezium Technologies) are clear: not a military robot; no weapons. Ever. But using the Sentinel, a police officer can remain in a car and use the public-address system to tell a driver to expect a visit from the robot. The robot trundles over and raises the tablet. The driver sees the police officer’s face; the police officer sees the driver and the car and can use simple programs to get a good look at license and registration. No sudden moves; no accidental gunfire.

Having participated in the university-wide Duke Start-up Challenge and the Pratt School’s DUhatch student start-up incubator, Tadepalli says “both programs have been wildly beneficial to us,” connecting them with mentors and encouragement.

Working directly with police departments, the pair learned they needed to reassure police officers that their idea was a tool, not a replacement for law-enforcement officers. From community groups they learned that a robot might seem even more intimidating than an officer, so they have been reassuring about the robot’s lack of weaponry and its simple job.

And they’re asking questions. “That’s an interesting side of it that we’re trying to be a part of,” Tadepalli says. The cheapest robots available to law enforcement now run in the $12,000 range and can’t do the job; they’re hoping the Sentinel can be cheaper, sturdier, and more useful, so that even small agencies can afford it. And the point isn’t to win grants or make products. The point, Tadepalli says, is simple: “Can this really make the difference we want?”