For more than five years, a team has been developing and testing control software under Alphabet's X moonshot factory, aimed at making industrial robots smarter and easier to use. Now the project has branched out into a new company called Intrinsic.
Alphabet has some impressive form in robotics of course, Google having acquired Boston Dynamics in 2013 before passing the torch to SoftBank in 2017 (which subsequently sold it to Hyundai). But Intrinsic appears to be keeping its hands away from the expensive realm of building its own hardware, and concentrating on developing software and AI tools that use sensor data from industrial robots to allow them to "sense, learn from, and quickly adapt to the real world."
According to a blog post from the company's new CEO, Wendy Tan-White, the idea is to leverage Alphabet's software development know-how to make the industrial robot workforce easier to use, more flexible and less expensive.
For the last five years or so, Tan-White's Intrinsic team has been looking at ways to give the kind of bots that assemble cars or consumer tech products on modern production lines the ability to "sense, learn, and automatically make adjustments as they're completing tasks, so they work in a wider range of settings and applications."
Developers have worked with outside partners in manufacturing, as well as the wider Alphabet community, to test the team's software creations, which are designed to train industrial robots using such tools as deep learning, motion planning, force control, automated perception, reinforcement learning, and simulations.
By way of example, the technique helped the team teach a robotic arm to plug three types of power connectors in the correct ports on a board in just two hours, whereas it may have taken skilled programmers hundreds of hours to generate the code necessary to teach the bot to perform this task. In another development, a team of bots worked together to assemble a piece of furniture. And for another trial, four bots used Intrinsic software to build wooden pods for an architecture project.
"None of this is realistic or affordable to automate today – and there are millions of other examples like this in businesses around the world," said Tan-White. "This all hints at the potential for Intrinsic’s software to radically reduce the time, cost, and complexity required to use industrial robots – and therefore their long-term potential to help with a much wider range of problems and drive up the diversity of goods that can be produced affordably and sustainably."
With the new company now launched, the team is on the lookout for partners in automotive, electronics and healthcare industries to take the project to the next level. It's also on the lookout for expert talent in software engineering, robotics and AI.
Source: Alphabet X