A Carnegie Mellon team recently developed software that teaches robots learn new tasks, simply by observing people doing the same tasks first. These robots have discovered how to open cabinets and remove garbage bags without direct instructions and could eventually learn how to get chores done just by watching YouTube videos. While the research is still preliminary, the software offers a glimpse into an uncertain future in which robots are more helpful companions in the home.
Tech companies have been teasing with jetsonslike future where robots will clean our counters and mop the floor. Dyson, a company known for its luxury vacuum cleaners, revealed in May that it is building a team to develop robots that can sort through the dishes and even clean under sofa cushions. Samsung presented last year that we will soon be able to have robot butlers zoom around the house, pick up dirty towels and pour glasses of wine. Still, aside from smart speakers and semi-automatic devices, home robots are hardly common in the average household at the moment. But the future of these devices — and what they might eventually do in our homes — is likely to take shape in the coming years.
“The idea is that you don’t have to wait for the robots to collect billions of data across numerous scenarios to learn something general and then be deployed,” Deepak Pathak, the Carnegie Mellon professor who worked on the project, told Recode. “It completely bypasses that process by placing a robot directly in homes, and helps them improve themselves in that environment through practice.”
Versions of home robots have been around for years and they are getting a lot more useful. Robot vacuum cleaners like Roomba, that’s about two decades oldhave evolved from relative simple automatic robots in artificially intelligent devices that work with smart speakers and record computer vision to study the rooms they are cleaning. The latest Roomba models can even travel to and from charging docks, where they empty trash into a box, all on their own. Amazon has built in a similar kind of navigation technology to build a security robot called Amazon Astro. Similar to Wall-E, this microwave-sized bot can roam around your house and record video when you’re not there. It also functions as a personal assistant that can recognize family members and follow you around.
Advances in AI have also led to a whole different class of robots that can perform more specialized tasks, such as: shovel snow, emptying the litter boxand cleaning grids and swimming pools. There are also social robots, designed to simulate companionship, set reminders and anticipate the schedules of the people who use them – an application especially useful for seniors. One such device, called ElliQ, has recently gone on sale in the US and New York State is already planning to… distribute 800 of these robots are among the state’s older residents. One ElliQ robot costs $250and then another $30 per month for an annual subscription to the robot’s content.
Most of these robots can’t accomplish much more than they’re explicitly designed for, so the idea of spending hundreds or even thousands of dollars on one of these devices can be unappealing. This is the problem that the Carnegie Mellon researchers tried to solve by designing their software called WHIRL, or In-the-Wild Human Imitating Robot Learning. WHIRL can be installed in any robot and adapts based on the physical capabilities of that particular device. After studying what the human is doing in his house, the robot tries to teach itself how to complete the same task, with whatever mechanical limbs it could be.
“Each job is unique, and we as humans can perform all those tasks,” Pathak, the Carnegie Mellon professor, told Recode. “Our robots are not capable of that at the moment. They are the opposite. They can only do one task in one environment.”
The most promising developments in AI-powered robotics have yet to hit the market. And many of the robots that consumers can buy still struggle with basic problems: Amazon’s Astro robot can move on flat surfaces, but can’t climb stairs, and sometimes struggles with navigation problems. Most home robots similarly lack the dexterity needed to grab and hold objects, which is a prerequisite for most chores. There is also the risk of the robot making a mess instead of cleaning one up. iRobot was famous for its Roomba. to update software after pet owners complained that the vacuum cleaners couldn’t see dog poop on the floor and would run over and smear it all over.
Home robots could get a boost as smart home technology takes off. Apple, Amazon, Samsung and Google are already working together on the matter, a common platform for smart home devices made by several companies. The hope is that technology can eventually control a fleet of our devices, which together can monitor security cameras, adjust the thermostat, and turn lights on and off. This idea that the whole house could become more autonomous does indeed coexist with a robobutler’s dream.
“You can almost imagine a Rosey The Robot scenario where you have this one super complex robot that can do it all. It can vacuum your floor, fold your laundry and do the dishes,” said Chris Jones, chief technology officer at iRobot, the company that makes the Roomba robovac “An alternative vision is actually more like the bridge in Star Trek. It’s a kind of environmental intelligence that puts together many devices in the home that essentially form one large distributed robot.”
Still, it’s not yet clear how home robots will ultimately fit into that picture, as many are still not that sophisticated and feel more like a gimmick than a genuinely helpful hand.
“With the success of things like Google and Alexa, which have pushed artificial intelligence into homes, maybe that broadens the gates for more robots. But I think robots are still tripping over certain physical things,” said Scott Midson, a liberal arts professor. arts at the University of Manchester “They’ve had all this success talking to us and learning our idiosyncrasies and our actions, but for the most part, robots are still learning the idiosyncrasies of our environment.”
Hopefully home robots will get better and our apartments and houses will start to look a bit more like the jetsons. After all, if the smart home becomes a reality for most people, a constellation of home robots, overseen by smart home technology, could take on all sorts of tasks we’d rather not do. This would give us humans a lot more time to do things we love, or at least things we enjoy much more than taking out the trash.
This story was first published in the Recode newsletter. Register here so you don’t miss the next one!