Cleaning, navigation: Here’s how a robot vacuum works

Robot vacuums are as useful as they are mysterious. Part of the mystique comes from the fact that robot vacuum manufacturers use acronyms and techno-jargon to describe their products' features.

In this article, we answer the question “How do robot vacuums work?” so you can get a sense of how they work and understand some of the buzzwords used about these automated floor cleaners.

Most importantly, this is how a robot vacuum cleans

There are many variations in robot vacuum design, but most follow a similar pattern. A good example is the iRobot Roomba i3+. It has a side- or edge-sweeping brush and a center brush roll, usually made of a rubber-like material, that helps clean away dust that has adhered to the floor.

Aside from brushes that sweep up dirt and debris, robot vacuums also have motors that create suction. Air gets pulled into the dustbin and then passed through a filter. Most robot vacuums have a single filter, but models like the iRobot j7+ or Shark RV1001AE have filters inside the robot vacuum as well as an additional one located in the self-emptying base.

Without navigation, robot vacuums wouldn’t be able to clean well or autonomously

The majority of robot vacuums on the market today use one or more of the following technologies:

• Lidar: This technology uses pulses of light to judge distance. Through multiple pulses, a robot vacuum can determine the shape of a room and any obstacles within it. Lidar is a cost-effective and reliable technology.

• vSLAM: Visual simultaneous localization and mapping—or vSLAM—is a camera-based method of robot navigation. By analyzing a succession of video frames, a robot can determine the distances to objects and their locations relative to them. Robot vacuums using vSLAM can “see” more of the room per sweep than models using Lidar.

• Object Recognition: This is an emerging technology in the robot vacuum world. With machine learning, robot vacuums can actually figure out what their camera is looking at and act accordingly. For example, a robot vacuum with object recognition would know to avoid power cables but also know to drive through a bed skirt.

Without one of these “seeing” technologies, your robot vacuum would be more likely to wander and clean the same spot repeatedly. This was a major complaint consumers had with the first generation of robot vacuums.