![WEARABLE_INSIGHT [FORUM]](https://wearableinsight.net/wp-content/uploads/2025/04/로고-3WEARABLE-INSIGHT1344x256.png)
Zero Velocity Updates (ZUPT) is a compensation technique to reduce the accumulated position error (drift) in an inertial navigation system (INS). It is mainly used when tracking movement based on accelerometer and gyroscope data, and is especially effective in walking, running, and punch tracking.
Summary of Principle
Inertial sensors integrate acceleration data over time, estimate velocity, and integrate velocity again to estimate position. However, this process accumulates even very small errors over time, leading to serious position error (drift).
To solve this problem, ZUPT compensates for sensor errors by utilizing the fact that “velocity must be 0” at moments when it is certain that there is no movement (stationary state).
How it works
1) Stationary state detection
The system detects changes in acceleration and angular velocity in real time and finds sections where the sensor is stationary and not moving. Example: The moment the foot touches the floor and stops, the moment the hand stops after a punch, etc.
2) Zero Velocity Update
Reduces the accumulated velocity and position errors by forcibly setting the velocity of the corresponding section to 0.
At this time, the Kalman Filter or Extended Kalman Filter (EKF) is used together to adjust the state estimation.
3) Improved position accuracy
Regular application of ZUPT suppresses the drift phenomenon of position estimation and improves the overall tracking accuracy.
Example of use in sports (e.g. boxing tracker)
- ZUPT can be applied at the moment when the hand stops before and after a punch
- When the arm stops after a punch or maintains a ready posture, the velocity is corrected to 0
- As a result, the accuracy of punch trajectory analysis is improved by reducing the position and velocity estimation errors.
Applications
ZUPT is particularly valuable in:
- Pedestrian navigation: Detecting when feet touch the ground during walking
- Vehicle tracking: Identifying stops at traffic lights or parking
- Robotics: Improving odometry accuracy during pause periods
- Military/tactical systems: Maintaining navigation accuracy during stationary reconnaissance
Benefits and Limitations
The main advantage is significant reduction in position drift that would otherwise accumulate rapidly in pure inertial navigation. However, ZUPT requires actual stationary periods to be effective and may struggle with very slow motion or vibrating platforms where true zero velocity is difficult to distinguish from sensor noise.
The technique is often combined with other navigation aids like GPS or visual odometry for more robust positioning systems.
Why is ZUPT important for sports and fitness wearables?
ZUPT is crucial because it helps maintain the accuracy of movement tracking in fitness wearables. By compensating for drift, it ensures that the data collected (like distance and speed) is more reliable and reflective of actual performance.
Can ZUPT be used in all types of wearables?
While ZUPT can be applied in many types of wearables, its effectiveness largely depends on the specific use case and the algorithms implemented. Devices designed for activities with distinct periods of inactivity are more likely to benefit from ZUPT.
How does ZUPT compare to other error correction techniques?
Compared to other techniques, ZUPT is particularly effective for activities with clear moments of zero velocity. Other techniques may focus on different aspects of error correction, such as filtering or sensor fusion, but ZUPT provides a specific correction method for drift during stationary phases.
