Robot eyes. That’s a terrifying sentence. But robotics vision is an immense area of research interest, and a key technological field in terms of building the future of a wide variety of devices. That’s why it’s very interesting that Dyson is putting a sizeable investment into robotics vision via a joint robotics lab being launched in collaboration with Imperial College London.
The investment is worth £5 million (or around $8 million U.S.) and covers a five-year period. The lab will be working on robotics vision systems that are designed to help the next-gen of robots not only see things the way that humans do, but also process that visual information in a manner that better approximates human understanding.
For those unfamiliar with the field, it covers a broad range of potential uses: A friend with a graduate degree in robotics vision engineering helps design systems for production lines that inspect the products being built for quality assurance purposes. Typically, these offer up margins of error that are tiny compared to the standards established by human inspectors.
Dyson is no stranger to conducting robotics research – the company has been exploring that area of interest for the past 15 years, according to the company. With Imperial College London specifically, it’s been working on developing systems that can view, interpret and “logically navigate” their surroundings. This applies to robotic vacuums in terms of Dyson’s business interest (the company mentions this product category specifically, so watch out Roomba) but it’s not their only goal in terms of applied robotics.
What this signals for Dyson is a graduation of sorts, as the company moves from thinking about robotics as an area of sustained but relatively light interest, into something it would like to ramp up on the production side. Hopefully at the end of these next five years, we’ll all be living with an army of Dyson home cleaning automatons, but at the very least we should see some advancements in terms of the ocular powers of our robotic friends.