Last week I wrote about a robot that does its best to avoid people when they’re angry. The video below shows another that’s pretty good at doing the opposite. Created by researchers at Brown University, it wheels along behind a person at a set distance, as if attached to an invisible rope.
The robot uses a camera to identify and follow the person in front of it. Usually this kind of system needs to be recalibrated for changes in lighting and camera movement.
The Brown team developed image-recognition software and combined it with a special depth-imaging camera. The camera uses infrared, allowing the robot to identify and focus on the silhouette of a user. “This allows us to reasonably segment people from background scenery with less dependence on lighting condition,” says Chad Jenkins, an assistant professor of computer science at Brown University who presented the work at this year’s Human-Robot Interaction conference in California. Even if multiple people pass in front of the robot, it still stays focused on its designated user. A controller can issue a set number of verbal or gestural commands, which the robot can recognize.
“It’s impressive work,” says Charlie Kemp, a professor at Georgia Tech University who works on robotic assistants and wasn’t involved with the research. “In general it’s very easy for robots to get confused, especially while they’re trying to keep up with a person.”
Challenges that remain include getting the robot to understand gestures from a person standing in profile, and refining the system to work in bright sunlight.