Robots for our households, like Rosie from The Jetsons or C3PO from Star Wars, are still a thing of our dreams. But what if our dreams could one day become reality? The Everyday Robot Project is working on just that – the project aims to create learning robots that assist humans in their everyday lives.
Led by Hans Peter Brondmo at X (formerly Google X), the team has observed that robots currently employed around the world can only operate in strict environments designed for them, the robots can’t adapt to new situations. The Everyday Robot Project is working on creating new forms of machine intelligence to allow robots to learn to adapt to new environments and situations, perfecting them to be used by the everyday person in their everyday life.
Brondmo recently shared the latest updates on their project. He tells that one task the Everyday Robot Project is currently teaching robots is how to sort waste. This means sorting out the recyclables into categories like plastic, paper, and compost.
Multiple technologies are being put to work here by the team, including sensors, cameras, simulation, reinforcement learning (learning by making mistakes and trying to improve), and collaborative learning (taking what’s been learned and sharing it with others). Their training takes place in two phases – virtually in a cloud simulator and then practically in the real world. Brondmo has explained that the real world practice by one robot can then be shared with the rest of the robots by putting the learned data back into the simulator.
By showing improvement, their tests have proved successful so far, enforcing the idea that learning technologies, as opposed to programmed technologies, may just be the future of robotics.