Robot vacuum cleaner Xiaomi: how to save the map of the apartment forever

Today's Xiaomi robot vacuum cleaners and the Mijia ecosystem can do wonders by creating accurate digital copies of your home. But many users are faced with the unfortunate situation: after successful navigation and cleaning, the map suddenly disappears or resets. It's not just a visual defect, it's the loss of all settings: no-go zones, routes and the sequence of cleaning rooms. Keeping a map is a critical process for automating everyday life.

The causes of data loss can be trivial, from unstable Wi-Fi to software firmware bugs. Often, tech owners do not know that a device requires a specific “learning” algorithm before fixing a layout in memory. Roborock, Vacuum Mop and other Android or proprietary OS models have their own nuances of working with LDS or vSLAM navigation modules.

In this guide, we will take a look at why the robot vacuum cleaner forgets its way home, how to do the first calibration, and what to do if the map is already damaged. You will learn about the hidden settings of the Mi Home app that will help stabilize navigation. Understanding the principles of lidar operation and route algorithms will avoid most errors in the future.

⚠️ Warning: Before setting up, make sure there are no extra items, wires or toys on the floor. Lidar may perceive hanging curtains or chair legs as obstacles, which will distort the geometry of the room.

Principles of navigation and map construction

To understand how to save a map, you need to understand how the robot sees it. Xiaomi devices use a laser rangefinder (LDS) or visual cameras (vSLAM). The laser sensor rotates, emitting beams and measuring the time of their return, creating a cloud of points. Cameras analyze the texture of the floor and the position of objects. The accuracy of construction depends on the purity of the sensors and the absence of interference within a radius of 1.5 meters from the base.

The scanning process is real-time. When you start cleaning for the first time, the robot doesn't just move randomly, it builds a vector model. If you interrupt the server connection or turn off the device, the data won't be written to permanent memory (NVRAM), which is why the first run should be uninterrupted (without interruptions).

The SLAM algorithm constantly checks the current position against the data already known, and if it gets lost, it starts building a new map or, in the worst case, resets the old one, often when there are abrupt changes in light or new large objects.

Step-by-step: first calibration and preservation

The right initial setup is the key to stable operation. Don't try to start floor washing or spot cleaning right away. First, the device must create a base layer. Follow these steps in strict order to fix the layout.

Set a charging station against a wall in a free space where the robot can approach it from different directions. Connect the device to a 2.4 GHz Wi-Fi network (5 GHz networks are often not supported or work unstable with IoT devices). Open the Mi Home or Roborock app.

Start the routine cleaning mode. At this point, it's critical that you don't touch the robot or move it manually. If it's stuck, you'd better release it and return it to the same place, but don't change its orientation abruptly. Once you complete the cycle and return to base, the map should automatically update in the app.

If the card appears but appears fragmented, repeat the procedure. Sometimes it takes 2-3 full cleaning cycles to form a stable digital twin of the apartment. After a successful scan, the system will suggest dividing the room into zones or renaming them.

Configure virtual walls and zones in the application

Once the map is saved, the fine-tuning phase begins, and the app allows you to create invisible barriers that the robot won't cross, which is useful for protecting long-pile carpets, where the pet bowl is, or areas with wires.

In the application interface, go to map editing mode, and you have the Virtual Wall (line) and the No Zone (square) tools available, and the line acts as a barrier that you can't cross, and the square is completely ignored by the robot when planning your route.

• 🚧 Virtual Wall: Draw lines along the rapids or boundaries of the carpet so that the robot doesn’t get stuck on the pile.
• 🚫 Zone of prohibition: use for places where children's toys are often standing or charging cables lie.
• 🏠 Rooms separation: The “Separate” function allows you to manually adjust the boundaries if the algorithm is mistaken and combines the kitchen with the living room.

⚠️ Warning: Don't create too many small no-go zones. Excessive virtual restrictions can overload the robot's processor when calculating the route, leading to getting stuck or skipping sections.

There's also a sequential cleaning feature, so you can set the order, first the kitchen, then the corridor, then the living room, and this is especially true for models with wet cleaning, so that you don't have to carry dirt around clean rooms.

💡

Use the Multi-storey Map function only if you really have multiple floors. For single-storey apartments, it is better to keep one main map to avoid confusion in the history of cleaning.

Wi-Fi and server problems: why the card is flying

One of the most common reasons Xiaomi doesn't save the card is because of an unstable network connection, and the robot uploads the card to the cloud for processing and storage, and if the router's connection to the router is interrupted when the cleanup is completed, the data may not sync.

Check your router's settings. Make sure that your IoT devices are static. IP-address or reserved DHCP. This will prevent a robot from getting a new address and losing communication with a previous session in the cloud.

If you have changed region (for example, from Russia to China or Europe), old maps may not be displayed, as they are stored on servers of a particular region, and in this case, card recovery will require returning to the original region of the account.

Failures after firmware updates and their elimination

Manufacturers regularly release software updates to improve navigation algorithms, but sometimes new firmware contains bugs that corrupt card files. If the robot starts to behave strangely after the update, first check the software version.

In rare cases, the update resets navigation settings. Don't panic. Often, it helps to just restart the device. Press the switch off button on the case (if any) or a combination of buttons to reboot the system. In Android, which many vacuum cleaners are based on, it's similar to rebooting a smartphone.

If the problem persists, try clearing the Mi Home app cache on your phone. Go to your phone's settings -> Apps -> Mi Home -> Memory -> Clear the cache. This won't delete your accounts, but update your local device data.

In extreme cases, if the card is physically damaged in the device's memory, a full reset will help. This is usually done by simultaneously clamping the power and return buttons to the base (or through the pin reset hole).

Maintenance of sensors for stable navigation

The technical condition of the sensors directly affects the quality of the map. Dust, wool and dirt are the main enemies of accurate navigation. Regular cleaning is not just hygiene, it is a necessity for the correct operation of SLAM algorithms.

First, clean the laser rangefinder's protective glass (Lidar). Even a thin layer of dust can scatter the beam, creating noise on the map. Wipe it with a dry microfiber. Also check the collision sensors ( bumper) and the elevation sensors (Cliff sensors) from below.

• 🧹 Laser Window: Wash with a Dry Wrap once a week.
• 👣 Floor sensors: remove dust from the bottom panel so the robot doesn't "think" that it fell down the stairs.
• 🔄 Wheels: Check the side brush and main wheel for coiled hair that can block rotation.

⚠️ Warning: Never use wet wipes or chemicals to wipe optical sensors. The fluid can leak in and cause the contacts to oxidize, causing permanent breakage.

Frequently Asked Questions (FAQ)

💡

The stability of the card depends not only on the software, but also on the physical position of the base and the purity of the optical sensors, regular maintenance is more important than frequent updates.