'Everyday Robots' Can Help You Deal With Everyday Tasks

Anil Singh


By using machine learning algorithms, engineers develop the robot, aiming to deal with an “unstructured world”.

You might have thought of a robot that can help you cut off the time on doing everyday tasks, and Alphabet – the parent company of Google – now introduces a new robot to make the idea go into real life. The so-called “Everyday Robot” boasts one of the most notable abilities: self-learning when facing new problems in the real world, rather than solving coded tasks.

By using machine learning algorithms, engineers develop the robot with the aim to deal with an “unstructured world” with various types of surroundings.

The robot has been under development of X – The Moonshot Factory, which has been the successor of Google X. The division focuses on making autonomous cars for Google’s projects, but now it appears to work on autonomous robots that can learn general purposes themselves. By using machine learning algorithms, engineers develop the robot with the aim to deal with an “unstructured world” with various types of surroundings. As such, the robot will take advantage of sensors and cameras to navigate as well as interact with objects through a robotic grasp.

In an experiment conducted an office environment, the robot successfully sorted trash with very high accuracy – 95 percent trash in the right places. After a bunch of tests, Everyday Robots have proved an outstanding performance when doing a new task without being programmed previously like other robotic machines. It’s said to get well with every stage, even the smallest individual movement. Everyday Robots now are only tasked with sorting waste, but the team expects to transform it into a multitasking robot.

Next Story