Alphabet deploys 100 'housekeeping' robots at Google's offices to run errands; Watch

In a significant development, Google's parent company Alphabet has been testing their prototype robots by making them perform small tasks autonomously in Google offices. In a bid to train robots to perform tasks in a complex environment like houses and offices, Alphabet has announced that its Everyday Robots’ project team, launched in 2019, within the Experimental X labs wing has moved some robots in Google’s office areas and are making them perform small, useful tasks like wiping tables, segregation waste, opening doors etc. 

Around a fleet of 100 robots are performing a range of errands autonomously in Google’s Bay Area Campuses, as the scientists have moved them out of the labs to make them useful in areas other than industrial zones. 

Google's parent company Alphabet tests 100 prototype robots to run errands

In a blog post talking about the experiment, the Chief Robot Officer of the Everybody Robots project, Hans Peter Brondmo, said, "We are now operating a fleet of more than 100 robot prototypes that are autonomously performing a range of useful tasks around our offices."

"The same robot that sorts trash can now be equipped with a squeegee to wipe tables and use the same gripper that grasps cups can learn to open doors," he added in the blog.

Alphabet has spent the last several years developing an integrated hardware and software technology for enabling robots to learn to transfer the knowledge from the virtual to the actual world. To understand and absorb the world around them, the robots are outfitted with a variety of cameras and sensors.

The robots have slowly gained a better grasp of the world around them and become more adept at executing ordinary activities using a combination of machine learning techniques such as reinforcement learning, collaborative learning, and learning from demonstration, according to the organization. With less than a day of real-world learning, a single robot can now learn how to do a hard task like opening doors with a 90% success rate, explained Brondmo.

"Over time, we will be expanding the types of tasks they are doing and the buildings where we operate and look forward to sharing updates from our journey over the coming months and years," said Brondmo in the blog. "Even more excitingly, we have shown that we can build on the algorithms and learnings from the door opening and apply them to a new task: straightening up chairs in our cafes," he added further.

Image: Twitter