Google deploys a fleet of robots to sort waste, wipe down tables and open doors
A fleet of over 100 Everyday robots from Alphabet’s X Division are deployed to perform some basic tasks, but with the opportunity to learn more.
Google could pour millions as he builds an error-correcting quantum computer, but he’s also experimenting with robots that are used on his campus to do something as simple as sorting through trash and wiping down cafeteria tables. . Google’s development team X has been interested in robotics for some time, but they have also done remarkable work on projects like Google Glass, a balloon internet system, the self-driving technology behind Waymo and a robotics project called Everyday Robot, among others.
Unlike Big Tech rivals or established names in robotics, Google hasn’t made remarkable progress or made wacky claims about its robotics ambitions. Facebook, for example, recently introduced a synthetic skin that can give robots a human-like tactile perception while handling delicate objects. CalTech’s LEO bipedal robot can walk, fly and even skateboard. Amazon has been making robots to deliver the products, while Apple’s disassembly robot named Daisy has been tearing up iPhones for recycling since 2018. But all great things start with one small step, and Google’s X Division abides by that philosophy.
Hans Peter BrÃ¸ndmo, Head of Robots at X Development, revealed that a fleet of more than 100 everyday robots would be deployed to Google’s Bay Area campuses to perform basic tasks like sorting through trash and wiping tables with a squeegee – and soon, opening the doors also. There’s nothing fancy like BostonDynamics’ Parkour robot, but the goal behind the X Division’s daily robots is a little more meaningful than that. The goal is to train robots to learn one task and apply those learnings to perform another task without requiring a ton of engineering upgrades or creating a vast library of code from scratch.
Learn new tasks based on existing skills
Instead of writing code for every task a robot can perform, Google relies on machine learning to have these robots learn to operate in unpredictable, unstructured spaces like an office or living room where they coexist with human being. The team applied ML-based methods such as reinforcement learning, collaborative learning, and demonstration learning to help the robot improve in its daily tasks by understanding the world in greater detail. ‘surrounded. The team pursues the natural advancement of abilities with coexistence in mind and not like the Tesla Bot. It takes a humanoid approach to performing tasks, but should be designed not to harm humans, even if it wanted to after software corruption.
BrÃ¸ndmo says the approach of building on existing abilities to learn something new has yielded positive results. For example, the algorithms and learnings behind the ability to open doors have been developed in such a way that the everyday robot can now also straighten chairs. These robots will soon be deployed across Google desks to automatically open the doors and check if it needs cleaning or if it is missing a few chairs. The end goal is to create a self-taught robot that can live sustainably instead of learning advanced social skills that allow it to choose between sabotaging or helping others.
Next: Electric Delivery Robots Reduce Emissions By Saving 280,000 Car Rides In A City
MBFFL: Why Whitney Thore Reflects On Her High School Days With Photos
About the Author