The principle of counting steps in Xiaomi Mi Band 3

Modern wearables have become an integral part of the lives of millions of users who want to control their physical activity. Xiaomi Mi Band 3 has established itself as one of the most popular trackers, but many owners are interested in how accurate its data and what mechanism is the basis of computing. Understanding the principles of the device helps to objectively assess the results of training and adjust the load.

It's based on a complex system of sensors and algorithms that analyze hand movements in three-dimensional space, and the accelerometer is the main component that captures acceleration and vibrations as you walk, and it transmits raw data to the processor for further mathematical processing, turning chaotic movements into understandable numbers on the screen.

It's important to understand that the device doesn't "see" your legs, it only interprets the wrist vibrations, which means that the filtering algorithm plays a crucial role in weeding out false positives that occur when you're working at a computer or driving a car, and the accuracy of the result depends on the calibration of the sensors and the individual gait of the individual.

The role of the accelerometer and gyroscope in tracking

The heart of the tracking system in the Mi Band 3 is a triaxial accelerometer, which is a miniature sensor that can detect changes in speed across three coordinate axes, allowing you to create a detailed profile of your hand moving through space. When you take a step, your hand makes a characteristic pendulum motion, which the sensor picks up as a burst of acceleration.

The gyroscope, while used primarily to determine screen orientation and sleep tracking, also contributes to the overall picture: It helps the system distinguish walking from other activities by analyzing the angular speed of the wrist. The combined operation of these components provides a more detailed picture of activity than using an accelerometer alone.

Modern algorithms analyze the frequency and amplitude of the vibrations. If the rhythm of the movements coincides with the typical rhythm of a person's step, the system counts the event. However, it is worth considering that the sensitivity of the sensor can vary, and sudden swings of the hands without walking can sometimes be mistaken by the system as distance traveled.

⚠️ Attention: Active hand gestures during a conversation or working with small details can lead to artificially overestimating the number of steps, as the accelerometer records sharp movements of the wrist.

Data processing algorithms and noise filtering

Raw sensor data is rarely perfect, so the Xiaomi Mi Band 3 uses sophisticated software filters to clear the signal from noise. Noise in this context refers to any movement that is not related to walking, such as brushing your teeth or typing on a keyboard. The algorithm compares the incoming signal with reference gait patterns.

The key is regularity of movement. Walking is a cyclical process, and the system looks for repetitive patterns. If the movements are chaotic or lack the pronounced frequency characteristic of the step, they are likely to be ignored, which minimizes the error of measurement in the home.

It also takes into account the duration of activity. Short-term bursts of movement lasting less than a few seconds are often not counted to avoid inflating statistics from random actions. The system waits for rhythm confirmation before starting an active countdown.

Effect of step length on distance calculation

Although step counting is based on mechanical movements, calculating the distance traveled in kilometers depends on the setting of the step length. By default, the Mi Band 3 uses averaged values based on height and weight, which you enter when you register with the Mi Fit application. However, the individual anatomy can differ significantly from the statistics.

If your step length does not match the algorithm, the distance displayed will be incorrect, even if the number of steps is calculated perfectly accurately. To work correctly, you should take a self-measurement. Go through a known distance (for example, 100 meters) with a normal step and count the number of steps, then divide the distance by the number of steps.

This will significantly improve the accuracy of the metrics displayed on the bracelet screen and in the application, and without this calibration, the distance error can be as high as 10-15%, which is critical for those who are preparing for marathons or following the standards.

💡

For maximum accuracy, measure the length of a step on a flat surface, walking 10 meters at a normal pace. Divide 1000 cm by the number of steps taken - this is your individual step length.

Comparison of accuracy with other devices

Users often wonder how the Xiaomi Mi Band 3 readings are realistic compared to other gadgets or mechanical meters. Tests show that when walking on a flat surface, the error is minimal and comparable to more expensive counterparts.

Below is a table showing the comparative performance of the tracker in different use cases:

As you can see from the data, the main problem for all trackers, including the Mi Band 3, is when the hand is fixed. If you push a cart in a supermarket or lead a child by the hand, the amplitude of wrist movements decreases, and the accelerometer can "sleep" part of the steps.

Factors that reduce the accuracy of measurements

There are a number of external and internal factors that can negatively affect the quality of tracking, and understanding these nuances will help you avoid disappointment from the data you get, primarily the way you wear the device.

If the bracelet is worn too loose, it can hang around on the wrist, creating vibrations that the accelerometer will perceive as steps. Conversely, too tight a clasp can limit the natural amplitude of movement of the hand, which will lead to undercounting. The optimal position is tight, but without compression, about 1-2 cm from the bone of the wrist.

Also worth mentioning is the effect of surface type: Walking on soft ground, sand or snow requires a lot of effort, but the amplitude of hand movements may be smaller because of the resistance of the medium, and in such conditions, the calculation algorithm may work less efficiently, skipping steps with low amplitude.

• 🏃‍♂️ Strange rhythm: A sudden change in walking pace or stopping in the middle of a step can knock down the synchronization algorithm.
• 🧤 Clothing: Wearing a bracelet over the sleeve of a winter jacket significantly reduces sensor sensitivity to bone movements.
• 🔋 Battery Charge: At a critically low charge level, some devices can go into power saving mode, reducing the frequency of sensor surveys.

⚠️ WARNING: Wearing a bracelet on long-sleeved clothing that falls on the wrist almost guarantees up to 30% of steps lost due to cloth quenching of vibrations.

Customization and calibration in the Mi Fit app

To get the most relevant statistics, you need to properly customize your user profile. Mi Fit (or Zepp Life) collects data on your height, weight, age and gender, and builds the initial model on these parameters, which is then adjusted during use.

The adjustment process is simple, but it requires care. Go to the device profile, select the settings of the body parameters and enter the current data. If your weight or height has changed, be sure to update this information, as it depends on the calculation of burned calories and step length.

You can also track your activity history by day or week, and if you look at the charts, you can see anomalies, for example, if you're on a rest day, you can double-check your sensitivity settings, or if you're looking at any activity that mimics walking.

Frequent questions about the tracker

💡

The accuracy of Xiaomi Mi Band 3 depends on the correctness of the anthropometric data entered and the density of the sensor adhesive to the skin of the wrist.