Smart navigation is the heart of any modern robot vacuum cleaner, especially in the Xiaomi ecosystem. Without an accurate map, the device will run chaoticly, bumping into obstacles and missing entire sections of the floor. Many users mistakenly believe that the map is built automatically on the first run, but for maximum efficiency, proper pre-preparation of the room and the gadget itself is required.
The process of creating a virtual apartment plan depends on the type of navigation installed in your model. Laser rangefinders (LDS) build a map faster and more accurately than visual navigation systems (VSLAM) or gyroscopic models. Understanding the principles of sensors will help you avoid the typical mistakes that lead to a loss of orientation of the robot in the middle of cleaning.
In this article, we will take a detailed look at the algorithm for creating the perfect map that the robot will use for years, and you will learn how to adjust boundaries, create no-go zones, and combine multiple floors into a single control system.
Preparation of premises and starting position
The first and most critical step is to physically prepare the space, and the robot must see the space as it looks in (constant use mode), but without any temporary obstacles, and remove wires, toys, clothes, and other small objects that could confuse the wheels or be mistaken by the sensor for an obstacle.
Pay special attention to the thresholds and transitions between rooms. If the height difference exceeds 2 cm, the robot may not overcome it, which will lead to a rupture of the map. In such cases, it is recommended to use special ramps or simply leave the door ajar for scanning, so that the device can move freely between zones.
Place the dock in the right place, it should be placed against a free wall, so that it has a meter and a half of free space left, right and front, and do not hide the base in niches or behind curtains, because the sensors must clearly see the charger for a successful return.
💡
The ideal place for a base is along a wall in a hallway or living room, where it will not interfere with the passage of people, but the robot will be able to freely approach it from any side.
Primary scanning and map construction
Once connected to Wi-Fi and added to the Mi Home app, the system will prompt you to start building a map, start a full cleaning or "Map" mode (if available in your model), at which point the robot will start a chaotic movement, exploring the perimeter of the room.
During the first pass, it is strictly forbidden to interfere with the operation of the device. Don't take the robot in your arms, don't move it or block the path. If the device is stuck, carefully release it and return it to the same place, but it is better to wait until it gets out of the predicament itself.
The duration of the initial scan depends on the area of the apartment and the number of rooms. For a standard two-bedroom apartment, the process takes from 20 to 40 minutes. At this time, you can watch the movement in real time through the application, but remember that active use of the phone screen can slow down the data transfer.
If the robot has lost orientation and started spinning in place, check the purity of the sensors. Wash the bottom cliffs (fall sensors) and side rangefinders with soft cloth. Dust on the optics is a common cause of failure in the construction of the geometry of the room.
Editing the map and dividing the rooms
When the initial scan is complete, you'll see the outline of your apartment in the app, and often the automatic division into rooms is incorrect: the kitchen can be combined with a corridor, and the living room divided into two parts, and editing tools can correct these inaccuracies.
Go to the map settings and select the "Separate" or "Merge" tool. Draw the separation line where walls or doorways should be, and the app will show you a yellow line that confirms your choice, so that in the future you can send the robot to clean only a particular room.
Room names also have implications for voice control through Alice or Google Assistant, renaming zones to their actual purpose: "Kitchen," "Bedroom," "Bathroom," and making it easier to create automation scenarios.
☑️ Checking the map
| Editing function | What is used for | Impact on cleaning |
|---|---|---|
| Separation (Split) | Dividing a large area into rooms | Allows you to clean only the kitchen or only the bedroom |
| Merge (Merge) | Connecting disparate pieces into one room | The robot will not stop between areas of the same room. |
| Rename | Assignment of names to zones | Needed for voice commands and statistics |
| Turning the map | Adjustment of North Orientation | Simplifies map navigation for the user |
Configuring virtual walls and forbidden areas
One of the most powerful features of a smart home is the ability to create invisible barriers, with virtual walls and No-Go Zones allowing you to limit your robot’s access to certain areas without having to install physical magnetic tapes.
A no-go zone is a square or a rectangle that a robot should never go into under any circumstances, and it's perfect for protecting animal water bowls, or piles of wires, or places where a fragile vase is standing, and a virtual wall is a line that the robot perceives as an insurmountable obstacle.
It's important to not overuse the number of zones. If you deny the robot access to half the apartment, the efficiency of cleaning will drop dramatically. Use this tool point-by-point. For example, create a "Couch Carpet" zone with dry cleaning only if you have a wet-wipe hybrid.
⚠️ Warning: Virtual walls are tied to the coordinates of the map. If you move furniture or carpet, be sure to update the position of the no-go zone in the application, otherwise the robot may try to drive on the carpet with a wet cloth.
Preservation and management of several floors
Owners of two-story homes often wonder how the robot remembers different layouts. Modern Xiaomi models (especially with LDS) are able to store up to 4-5 different cards in memory.
When you move the robot to another floor, it's best to manually start building a new map or select "Save as a new map" in the settings. Don't let the device try to match the new floor to the previous one, this will cause mush from the data and loss of orientation.
For each floor, create a separate map with a unique name, such as "1 floor", "2 floor", "Mansarda", which allows you to see the history of cleaning for each level separately and customize individual scenarios.
What to do if the robot loses the map?
Typical errors and their solution
Even if you follow all the instructions, there can be problems, most often due to external factors that the algorithm cannot predict, and understanding the causes of failures will help quickly restore the system to work.
The most common mistake is full-length mirrors and glass doors, where the laser beam passes through the glass or is reflected unpredictably, creating phantom walls or tunnels, and in such cases it is recommended to temporarily seal the bottom of the mirrors with paper or install physical barriers in front of the glass doors.
There are also problems with poor lighting in camera models (VSLAM), where if the room is dark, the robot will not be able to read visual anchors and will get lost, and it is critical for these models to turn on the lights or have good lighting at night.
💡
The stability of the map depends on the immutability of the landscape: try not to change the arrangement of large furniture often, so that the robot does not lose its binding to the coordinates.