The July 2022 issue of IEEE Spectrum is here!

Close bar

Dyson's Robot Vacuum Has 360-Degree Camera, Tank Treads, Cyclone Suction

Dyson spent 16 years and $47 million to develop this robot vacuum, but it's one of the most advanced that we've ever seen

4 min read
Dyson's Robot Vacuum Has 360-Degree Camera, Tank Treads, Cyclone Suction
Photo: Dyson

All of our wild speculation about Dyson's mystery video (well, maybe not all) has just been confirmed with the announcement this morning of the Dyson 360 Eye robot vacuum cleaner. If it looks familiar, it's because we got some of our guesses exactly right: the new robot looks just like the patent image we posted, and it's got a vision-based localization system that relies on a panoramic camera. Even those tank treads that we thought were a joke? Totally not a joke.

Here's what more than 200 Dyson engineers spent $47 million over the last 16 years creating:  

The 360 Eye is one of the most advanced robot vacuums that we've ever seen, especially when compared with the myriad of Asian knockoff Roombas with gimmicky features that don't actually improve cleaning. Here's what Dyson says is new and amazing in the new vac, hot off the press release:

Vision and Sensing

The Dyson 360 Eye vacuum sees its environment using a unique 360° vision system [which is sensitive into the infrared], and then builds a detailed floor plan [based on high contrast visual features] to intelligently and systematically navigate around a room and tracks its position [SLAM, in other words].
 
Infrared sensors work in conjunction with a lens on the top of the machine that houses a 360° panoramic camera. The camera can see all the room at once so the machine can accurately triangulate its position. It then uses landmarks within the images to establish how it has moved between each frame and update its model of the environment accordingly.
 
The camera in the Dyson 360 Eye robot vacuum takes up to 30 frames per secondenabling the machine to effectively interpret its surroundings. Because the shutter speed of the camera matches the machine’s speed of travel, its position is always accurate to within mm, so it knows exactly where it is in the room, where it is yet to clean and uses infrared sensors to detect where obstacles lie.
 
Algebra, probability theory, geometry and trigonometry combine to create the Dyson 360 Eye robot’s vision of the world. It took over 100,000 hours from a team of 31 robotic software engineers to create the navigation system.

Movement

Like an all-terrain vehicle, continuous tank tracks enable the Dyson 360 Eye robot vacuum to maintain speed and direction across all floor types, and over small obstacles [like socks].

Tracks were tested against a rolling road with steps to climb over. This tested the suspension for more than 342,000 bump strips over the span of 1,000 hours. The machine travelled more than 670 miles during rolling road testing.

Cleaning

Powered by the energy efficient V2 Dyson digital motor [which spins at up to 78,000 RPM], the Dyson 360 Eye vacuum uses Radial Root Cyclone technology to effectively separate dust and dirt, capturing particles down to 0.5 microns600 times smaller than the period at the end of this sentence. Dyson 360 Eye vacuum cleans better than any other robot vacuum cleaner first time around.

The brush bar on the Dyson 360 Eye vacuum cleaner extends to the full width of the machine, meaning that it doesn’t rely on side sweepers to reach the edges of the room. It uses patented carbon fiber brush bar technology to remove fine dust on hard floors, and stiff nylon bristles to agitate and clean carpets.

Wireless

A new Dyson-developed app for iOS and Android will let Dyson owners keep on top of the cleaningeven if they are out of the house. They can even schedule the machine if out of the countryreturning from vacation to a clean home. Set up a single clean or a recurrent cleaning schedule at the flick of a button.

Other relevant facts: the dust bin, which you have to empty yourself, holds up to 0.4 liter of ick. The robot runs on a lithium battery which gives it a meager 20-30 minutes of run time, but it'll return to its dock to recharge, and then resume its cleaning pattern where it left off.

Dyson, like every other company that manufactures a robot vacuum with localization capability, says that the 360 Eye cleans efficiently by mapping the room and then planning a systematic route that avoids making multiple passes. iRobot, on the other hand, argues that you clean significantly better by making multiple passes from different directions. I tend to believe iRobot on this one (based on reviews I've conducted of various robot vacuums), but Dyson vacuums are famous for their ludicrously powerful motors, which may offset that.

There are, however, other issues to consider when it comes to robotic vacuums. With its Roomba 800 series, iRobot completely reengineered the cleaning system, creating a brushless design that "is virtually maintenance free and provides superior vacuuming performance on all areas of the floor, including under and around couches, beds, and other furniture," iRobot CEO Colin Angle said in a statement provided to IEEE Spectrum. In contrast, the Dyson robot uses a brush bar with bristles (which users will find maddening to clean), and it's too tall to go under furniture.

We'll have to wait until we can test one out to know for sure. It's also worth mentioning that some 360 Eye press materials suggest that the vacuum can identify what kind of surface it's on, and will prioritize cleaning "easier" surfaces first. 

We don't have information from Dyson as to how much the 360 Eye will cost, but according to some rumors, it might be US $1,650. This, obviously, is a lot, but if it really can take over as a vacuum replacement robot (as opposed to a vacuum reduction robot, like a Roomba), it could conceivably be worth it.

The Dyson 360 Eye launches in Japan next spring (that's 2015), with the rest of the world to get it later next year.

[ Dyson 360 Eye ]

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

“I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

This article is part of our special report on AI, “The Great AI Reckoning.”

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓Show less