Often when thinking about the future we imagine a world in which robots will replace humans in the most dangerous, tiring and tedious jobs. And if cleaning and household chores are high on the list for you, perhaps it’s time to celebrate: a robotic maid, in fact, could see the light sooner than you would have hoped. Thanks to three young engineers from Stanford University, who, working with colleagues from Google’s Deep Mind program, developed the Aloha mobile platform, a self-propelled robot capable of learning and perfectly replicating a wide variety of daily tasks, which ranging from household cleaning to preparing meals.
The capabilities of the new robot are described for now in an article available in pre-print, i.e. shared by the authors before being subjected to peer review which guarantees the quality of the studies published in scientific journals. The videos that show the capabilities of mobile Haloa, however, speak for themselves: we see him vacuuming, cooking, washing dishes and rinsing dishes, tidying up furniture and pots, with a naturalness that borders on that of a human being.
How does he do it? The merit – explain its inventors – is due to a combination of engineering and programming precautions. First of all, the robot is a self-propelled platform equipped with two mechanical arms, therefore more capable of replicating the actions of human beings compared to static models or models equipped with only one mobile arm. Furthermore, the movements it makes are not programmed step by step by its developers, but are learned by the robot by exploiting the potential offered by artificial intelligence.
The prototype is in fact equipped with a teleoperation system (i.e. remote control) which allows a human operator to pilot the robot like a kind of remote-controlled puppet, to teach its artificial intelligence how to complete new tasks and activities. Once the inputs provided by the operator in the training phase have been recorded, the robot’s machine learning algorithms train further by comparing what they have learned with the data obtained in the past from other static robotic platforms, ultimately “learning” how to perform the same operation autonomously , achieving almost perfect accuracy.
As we were saying, for now no external expert has validated the claims of the three researchers who developed the Aloha mobile, and the robot is still a prototype that will probably require several improvements before it can actually set foot in our homes. However, it has a fundamental strong point: the price. The entire robotic system was in fact developed by spending just 32 thousand dollars, a fraction of what other similar robotic devices cost. This will allow many laboratories to replicate what was done by Stanford researchers, almost certainly giving a notable boost to research in the field of home robotics (and beyond). And at the same time, it makes the commercial development of such a robotic platform more than plausible, at a price that is also affordable for domestic consumers.