Modern robot vacuum cleaners from Xiaomi and their sub-brands like Roborock or Vacuum are complex navigation systems that require initial calibration to work efficiently. Wet cleaning adds an additional level of complexity, as the device must clearly understand where the dry carpet ends and the laminate or tiles begin. Without a properly constructed digital map, the robot will move chaoticly, skipping corners or, worse, flooding carpets with water from the tank.
The process of building a map in the Mi Home or Xiaomi Home application is based on technology LDS-It's a scan that uses a laser rangefinder to create an accurate plan of the room, so you don't have to draw a diagram by hand, you just have to give the device freedom of movement. However, automatic mode does not always take into account individual layout features, such as rags, dark carpets or areas with a lot of wires, so manual adjustment of the map is a must to achieve the ideal result.
In this guide, we’ll take a look at everything from first calibration to complex no-go scenarios. Adjusting virtual walls and dividing rooms will allow you to manage your cleaning by voice or schedule without worrying about the safety of furniture and flooring. A deep understanding of the logic of the navigation module will help you avoid the typical mistakes when a robot gets stuck or loses orientation in space.
Preparation of premises and initial calibration
Before you start the scanning process, you need to physically prepare the space. The robot must have unimpeded access to all corners of the room, so it is recommended to temporarily remove from the floor toys, shoes, dangling curtains and long wires from the chargers. LDS-The sensor located on the top cover is sensitive to high obstacles, but can get entangled in a tangle of wires on the floor, which will lead to navigation error and interruption of map construction.
Make sure the dock is mounted against the wall in an open space, not in a niche or corner formed by furniture.The distance to the side obstacles should be at least 50 cm so that the robot can freely maneuver when leaving and entering the base.If the starting position is chosen unsuccessfully, the algorithm may misidentify the center of the map or create several unrelated islands, which will require a complete reset and repetition of the procedure.
☑️ Preparations for first launch
It's also important to assess the condition of the floor. Dark, especially black, carpets can absorb the signal of the drop sensors, which will make the robot consider them a cliff and avoid these areas. In such cases, it is recommended to either tape the drop sensors (if the model allows), or temporarily remove such carpets during calibration. Once you create the main map, you can manually designate these areas as forbidden or adjust the mode of ignoring them.
⚠️ Warning: Don't try to carry the robot while you're working or building a map. If the device gets stuck, first release it, then return it to base to restart the task.
Automatic map construction process
The mapping process is started via the Mi Home app. After connecting the vacuum cleaner to the 2.4GHz Wi-Fi network (importantly, 5GHz networks are not supported by most models), go to the device management interface. In the lower menu or in the settings, select the option “Map” or “Maps Editing”. For the first calibration, you must select the “Full Cleanup” mode or a special “Mapping” mode, if it is available in your firmware version.
The robot will start moving around the perimeter of the room, building the outline of the walls, and then fill the interior with zigzag movements, during which time the diagram will be displayed on the smartphone screen in real time, the process can take from 15 to 40 minutes depending on the area of the room, at which point it is critical to ensure a stable Internet connection, since the geometry of the room is transmitted to the server and processed in the cloud.
If you open a door to a new room during the scanning process, the robot will likely go in and expand the map. However, for better accuracy, it is recommended to scan the apartment in stages, opening access to new rooms only after the previous ones have been successfully completed, which will allow the SLAM algorithm (Simultaneous Localization and Mapping) to calculate coordinates more accurately and avoid layering.
In some cases, especially in large apartments or complex spaces, the robot may lose contact with the base or forget where it is. If a loss message appears on the screen, don't panic. Return the device to the base manually, press the pause button and then resume cleaning. The system will try to re-determine its location relative to the already built section of the map.
Editing the map and dividing into rooms
Once you've finished scanning, you'll get a single contour map of the entire area. To effectively manage wet cleaning, you need to divide that area into logical zones: kitchen, bedroom, living room. In editing mode (usually a pencil icon or "Separation"), you can draw lines manually where the algorithm has not recognized walls or doorways.
Use the "Separate" tool to draw a line along walls or imaginary boundaries. The app will automatically suggest optimal separation options, but manual adjustment often gives the best result. For example, if you have a living room with no obvious partitions, you can mark the border yourself so that you can wash the floor only in the cooking area without touching the rest area with the carpet.
| Tool. | Function | When to use |
|---|---|---|
| Split up | Divide one room into two | Studios, kitchens and living rooms |
| Unite | Connects neighbouring areas | If the corridor is broken apart |
| Name. | Rename the room | For voice control |
| Order. | Room numbering | For consistent cleaning |
Remember to rename rooms according to their purpose. It's not just a cosmetic function: "Kitchen," "Bathroom," "Bedroom" are used by voice assistants and smart home scripts. If you say "clean up in the kitchen," the robot needs to know exactly which area on the map matches that query. Naming errors can cause wet cleaning to start in the wrong place.
The Secret to Accurate Separation
Configuring virtual walls and forbidden areas
Virtual walls and no-go areas are key to wet-cleaning models. Unlike the physical magnetic tapes used in older models, the boundaries are set programmatically directly on the map. There are two main types of constraints: Virtual walls (the robot will not cross the line in either dry or wet mode) and non-cleaning zones (ignored only when wet actively clean).
To protect carpets and electronics, you need to use No-mop zones, and you draw a rectangle or a polygon on top of the carpet, and when you install a water module, the robot will automatically bypass the area, even if you've started cleaning the entire apartment, so you don't lift the carpets or move the furniture every time you do the floor cleaning.
Virtual walls are more convenient to use to restrict access to certain rooms or dangerous areas, such as a computer wiring congestion or an open balcony. You can create up to 10 of these zones on a single map. When drawing, try to step back from a real obstacle by 10-20 cm, because the robot is physically sized and inertia, and can hit the edge of the forbidden zone with a bumper.
⚠️ Warning: Virtual walls and no-go zones are map-bound, and if you move the robot to another apartment or room that has already been calibrated, the old restrictions may not work or may block cleaning in a new location.
Wet cleaning scenarios and mode selection
The map setting is closely related to cleaning mode choices, and in the Mi Home app, you can set individual settings for each room, for example, for the kitchen, you can set high suction intensity and maximum water supply, and for the bedroom, you can set quiet mode and minimal hydration, this is achieved through the room cleaning settings function.
Water supply regimes are usually divided into three levels: low, medium and high. Low is suitable for daily maintenance cleaning and laminate, which is afraid of moisture. High is designed for tile in the bathroom or kitchen, where there is persistent contamination.
💡
Use the Cleaning Order function to allow the robot to first vacuum all dry areas and then, after replacing the tank, move to wet cleaning, which will prevent dust from smearing on the wet floor.
You should also pay attention to the Y-trend function (if supported by the model). In wet cleaning mode, the robot can move not only straight lines, but also reciprocal movements, simulating manual washing. This greatly improves the quality of cleaning, but increases the time of operation. You can find the setting for this option in the Cleaning Modes -> Wet Cleaning menu.
Keeping maps and working with multiple floors
Modern Xiaomi and Roborock models allow you to save up to 4 different maps. This is ideal for multi-storey houses or apartments with complex layouts where separation is required. After building the first map, the system will suggest saving it. To add the second floor, just take the robot to the next level, start building a new map and save it under a new name (for example, 1 floor, 2 floor, cottage).
When you move a robot between floors, it's important to keep it from building a map on the way, and the best way to do that is to turn it off, move it to a new floor, put it on the base, and then turn it on, and if the robot starts building a map in the hallway or on the stairs, it can lead to levels mixing, and you can set up the no-go zones and schedules in each card setting.
If you plan to move the base frequently (like carrying a robot from room to room), you won’t be able to use the multi-card feature, so the robot will try to complete a single map every time or lose location. For such scenarios, it’s better to use “no card” mode (if available) or just start local cleaning, but full navigation will not work.
💡
Keeping multiple cards allows you to use one robot for different floors or apartments, while keeping individual settings of restricted areas and schedules for each room.