The Everyday Robot project performed by Alphabet's X group is making robots that can safely operate in human environments, where things change every day, people show up unexpectedly, and obstacles appear out of nowhere. In order for robots to be useful day to day, they need to understand and make sense of the spaces where we live and work, and adapt to them as they gain experience. This requires new forms of machine intelligence.
The challenge
Building robots that can operate autonomously in unstructured human environments, like our homes and offices, is a complex, unsolved problem. It requires tackling and integrating some of the hardest hardware and software challenges in the field of robotics today. The Everyday Robot project is building a new type of learning robot—one that can eventually learn to help everyone, every day.
Learning
The goal is to build robots that can learn from human demonstration, the experiences of other robots, and even from simulation in the Cloud. The Everyday Robot project has collaborated with Google AI to explore new ways of unlocking these challenges. High-fidelity simulation enables everyday robots to learn faster through practice in virtual everyday environments.
Robot systems designed for an unstructured world
To operate in our unstructured world they make use of cameras in the robot’s head and sophisticated machine learning models help our robots see and understand the world. The robot is designed to grasp, move, and interact with all kinds of everyday objects through a sophisticated arm. The robot uses data from its sensors to create an understanding of what it is seeing, hearing and where it is in the world – allowing it to safely perform useful tasks among people in everyday environments.