Gyro Drift Correction for An Indirect Kalman Filter Based Sensor Fusion Driver

Area of Science:

• Mobile computing and sensor technology
• Signal processing and data fusion

Background:

• Mobile applications rely on multi-sensing data (e.g., navigation, fitness trackers, VR games).
• Gyroscope and accelerometer accuracy are critical for user experience (UX).
• Gyro drift in indirect Kalman filter-based sensor fusion degrades orientation computation accuracy.

Purpose of the Study:

• To propose a novel mechanism for resolving gyroscope drift in sensor fusion.
• To enhance the accuracy of orientation computations in indirect Kalman filters.
• To provide a solution compatible with legacy applications without reprogramming.

Main Methods:

• Developed a mechanism targeting external feedback loops and non-gyro error elements in Kalman filter state vectors.
• Implemented the mechanism at the device-driver layer for low latency and transparency.
• Validated the mechanism's effectiveness through experimental analysis.

Main Results:

• Significantly reduced root mean square errors (RMSE) from 6.3 × 10⁻¹ to 5.3 × 10⁻⁷.
• Demonstrated improved accuracy in orientation computations.
• Confirmed compatibility with existing applications, requiring no reprogramming.

Conclusions:

• The proposed mechanism effectively resolves gyroscope drift in sensor fusion.
• The solution enhances mobile application accuracy and user experience.
• Device-driver layer implementation ensures broad applicability and ease of integration.