What is Clean by Object?
As a new feature with app version 5.0, you will be able to get clean zone suggestions based on robot identified objects in your home. These zone recommendations are automatically created, sized, and named around specific, detected objects. Examples of objects that will be suggested are couch, dining table, and counter. This feature gives you the option to accept or reject the suggestions. When accepting the suggestion, you also have the ability to adjust the name, size, and location of the zone. If a recommendation is denied, you can provide feedback on why the suggestion was denied. This feature is available only for the smart mapping robots and is also available with voice-assistant integrations.
How will I know or be notified that an object is detected?
When an object is detected as a clean zone recommendation, a message will appear under “Messages” within your app.
When you click on a message/recommendation, you are taken to a map view where the clean zone suggestion will appear as a highlighted blue rectangle on the map, with additional information provided. You are then walked through the process of confirming the object and its position and can edit the size, name, and position of the recommended clean zone themselves.
It is important to note that in order to receive recommendations, you must first agree to allow the App to send notifications and tips (specifically the option of ‘Personalized tips and reminders’ must be toggled on.
What if the object which is recommended is wrong?
We refer to this as a false positive – you receive a zone recommendation, but it is incorrect in size or location. An example of this is a recommendation for a couch in the bathroom, or a dining table in the bedroom. In this case, just reject the suggestion or accept it and edit the name of the object.
What if I’m not recommended a zone/object at all?
We refer to this as a false negative – you do not receive a zone recommendation at all. An example of this is – there is a couch in the living room but we don’t make a recommendation for that couch. In this case, you still would be able to manually place a zone in the area you want to target a cleaning job. Given the unique nature of customers’ homes and furniture, it is not possible for the robot to be correct all the time.
How do the new Clean Zones recognize objects?
Using state-of-the-art robot AI, Roomba robot vacuums and Braava jet robot mops with Imprint™ Smart Mapping use machine learning to automatically detect and proactively recommend Clean Zones around specific objects, like couches, tables and kitchen counters. This allows for targeted cleaning in specific areas or around objects that attract the most common messes. Simply use the app or request to a voice assistant device, “Roomba, clean in front of the couch,” and Roomba knows right where to go.
But how does it know what objects are?
Software powered by iRobot Genius is trained with thousands of images of common household objects, including couches, tables and chairs. Using machine learning, this allows our robots to automatically detect objects within the home and proactively recommend Clean Zones around these specific objects.
Is Roomba taking pictures of my home and identifying objects that way?
No. Software powered by iRobot Genius is trained with thousands of images of common household objects, including couches, tables and chairs. Using machine learning, and sensors that recognize distinct patterns throughout the home, our robots can automatically detect objects and proactively recommend Clean Zones around these specific objects. All of the processing takes place onboard the robot, and no images are sent to the cloud.
Can a user opt-out of some of these features, like Clean Zone/Object recommendations that detect objects in the house?
Clean Zone recommendations are a part of the Imprint™ Smart Mapping suite. Users can turn off this feature by toggling the appropriate button in the iRobot Home App settings that instruct the app to not send data. However, customers will lose functionality of Imprint™ Smart Mapping.
