We're still trying to get robots to walk through difficult terrain (or gallop like ostriches), but sooner or later they'll need to master another important skill: Not getting in our way. MIT researchers have built a wheeled 'bot that can safely navigate through crowds of people.
It's not just staying out from underfoot: The robot is programmed to follow subtle social rules too, like passing on the right and speedwalking past people on their left. The 'bot uses a suite of RGB cameras and a LiDAR array paired with face-detection tech to recognize humans and situate itself accordingly. Then it anticipates where they're going and, polite little robot that it is, weaves patiently around pedestrians so as not to get in their way.
That would make it ideal for busy areas with a lot of foot traffic, like malls or hospitals. The robot's goal isn't just to stay in its lane but to prioritize human walkers -- think of it as a prototype robot butler that's hyperaware of social cues. It runs a proprietary algorithm to navigate and respect particular traffic rules (passing on the right, which is the preference in the US), but those could easily be reprogrammed to adapt the bot to other cultures' walking norms. The MIT team, whose research was funded by Ford, will present their findings in a paper at the IEEE Conference on Intelligent Robots and Systems next month.
"The knock on robots in real situations is that they might be too cautious or aggressive," The paper's co-author Michael Everett toldMIT News. "People don't find them to fit into the socially accepted rules, like giving people enough space or driving at acceptable speeds, and they get more in the way than they help."
Sure, there are more pressing navigational priorities when some of our robots can't help throwing themselves into fountains. But respecting the subtleties of how we walk around and between other humans is downright civilized. The next time someone cluelessly stops moving in front of you on the sidewalk or passes on the wrong side, you can say, 'A robot would have done better.'