The new Roomba i7 + has persistent maps, selective room cleaning and automatic dirt removal

The Clean Base is cool enough for us to write an article about it that you can read here, but basically it's a docking station with an internal vacuum that can access the i7 + trash via a special port below, then transfer the contents of the trash into the dock's own trash, which is large enough to hold about 30 Roombafull.

Although the Clean Base was a complete surprise, which was not a total surprise, it is that the i7 + now has persistent cards. Just like the previous generation of high-end Roombas, the i7 uses a top-mounted camera that collects feature data from your home to generate a map via VSLAM (visual simultaneous localization and mapping). Earlier, Roombas created a new card each cleaning cycle, but cards created by the i7 + persist between cleaning cycles. In other words, the robot remembers the layout of your floor.

Since the cards stay around, you can do useful things to them with the iRobot app. This useful thing is currently the ability to tell the robot where the different parts of your house are, then clean it to clean only these parts: "Robot, go clean the kitchen." In practice, the robot will perform an exploratory cleanup. and then show you a map of your place with suggestions on how he thinks your rooms should be divided. You can then adjust (add, delete, reposition) the delineations and assign the labels of your choice to the parts. Each robot can memorize cards for 10 different floors, which is ideal if you live in a 10-story mansion like I do not, and the VSLAM system makes it particularly resistant to normal environmental changes such as moving furniture.

Now, not to brag or anything, but we have seen it all, because at IROS 2017 in Vancouver, iRobot researchers have presented an article on RoomSeg, an algorithm that automatically segments the parts of the room. a map created by a robot. In addition to allowing the cleaning of specific parts, the document also included data on the difference between the pre-existing card and the segmentation of the part on the efficiency of a cleaning robot:

While conventional conventional cleaning required a total of 2.32 hours, including 0.52 hours for the route and 1.8 hours for cleaning, room-by-room cleaning required only 1.9 hours, of which 0.3 hours for tracking and 1.6 hours for cleaning. Overall, the total execution time fell to 82%, the trajectory to 63% and 66%.

The i7 + also features a lot of small improvements, or maybe not so little – one of them, which I personally appreciate, is that the robot can now make the difference between the black carpet and the infinite void. In previous generations of Roomba, infrared cliff sensors could be mistaken for an infrared-absorbing black mat. This problem has now been solved, we are told, and there has been a lot of rejoicing and suppression of insulating tape and foil.

Efficiency is certainly important, but you may not notice it if you let your Roomba do its work while you're not there. The i7 + is the most autonomous Roomba ever combined with programming and database, because it can run for long periods of time without you having to manage it. iRobot says that your floors will always be cleaner. As Colin Angle, CEO of iRobot points out, in the press release, "This robot is up to the original vision we had to build a robot vacuum almost 20 years ago."

At $ 950 US for the i7 + with the Clean Base dock, it's a very expensive robot. If you want the mapping features but not the automatic scrapping, iRobot sells the same model (called just i7) for $ 699. And if you just want a robot vacuum cleaner and none of cartography and other bells and whistles, iRobot also announces today the Roomba e5, which costs 450 dollars. For people who live in a relatively small place and have little patience to handle a vacuum that will do nothing of sophisticated navigation and will not be able to empty, we are obliged to report it (As usual, your Soils will be almost as clean – an additional $ 500 for the i7 + are (almost) all useful but unnecessary features, as well as practical benefits.

That said, we are very excited about all persistent cards for home robots, and the i7 is the first iRobot model to exploit this capability. What we see now from iRobot (and also from Neato Robotics) is just the beginning, and these robots will now be smarter and more useful. For example, the i7 (like other Roombas) is able to detect which parts of your floor are the dirtiest while listening to the amount of things it picks up. As he locates and memorizes cards now, he should be able, over time, to create a map of places where dirt tends to accumulate in your home. At some point, he should be able to learn independently how to clean the dirtiest places most frequently, so you will not have to tell when and where to clean. Maybe it'll go under the couch once every few weeks, but take a walk around your door every day or two. And there is really no reason to think that iRobot could not implement this particular feature like tomorrow.

Beyond the obvious things like this one, just look at the interviews we did with Colin Angle to see where things are going and the iRobot team told us that we could 'wait a lot' interesting things in the short term. "The i7 is designed for the future". We do not know exactly how iRobot will be in the lead, but we're excited to know that.