Getting around in a public place isn't just about not walking into other people or avoiding collisions with pillars and park benches. When we walk, or just stand around in public, there are all kinds of unwritten rules we follow. And Stanford's Jackrabbot is a robot designed to learn these rules and to obey them.
Most of what we do in public, we do without thinking. For instance, when you stop to do something, you instinctively move over to the side of the sidewalk to get out of the way of others (or you should, anyway). And when you meet someone coming in the other direction, you both move around each other. You don't just stop and wait for the other person to do something, like a robot might do—in fact, if you did, it would seem aggressive.
And, unlike on the road, there's no rule book. Everything is done according to etiquette learned subconsciously, over years. The Jackrabbot is attempting to model these rules, effectively teaching itself how to behave around people.
The goal is to eventually build robots that can work alongside humans "in dynamic crowded environments such as terminals, malls, or campuses," says Stanford's project team, at the Computational Vision and Geometry Lab. As more jobs are taken over by robots, their ability to get along with people will become more important. As part of its training, the Jackrabbot is delivering mail, food, tools, and goods on the Stanford campus.
The training, which will teach the robot when to do things like taking its turn when waiting in line to, say, get to a busy exit, should help future robots blend into crowded areas. But the technology could also help elsewhere. Driverless cars, for example, could learn better how to interact with human drivers on the roads, perhaps becoming a little less meek in situations where people currently take advantage of autonomous cars' excessive politeness. And smooth crowd-navigation could also help visually-impaired people to better navigate public spaces, improving on the classic white stick.
The Jackrabbot is designed to be unthreatening. "We tried to make it cute, so that people will be happy to share their space with this robot," Stanford's Alex Alahi told the Bay Area's KCBS news.
[All Images: via Stanford Computational Geometry & Vision Lab]
Have something to say about this article? You can email us and let us know. If it's interesting and thoughtful, we may publish your response.