Efficient predictive estimator for holdover in GPS-based clock synchronization.

Area of Science:

• Electrical Engineering
• Signal Processing
• Metrology

Background:

• Accurate clock synchronization is crucial for Global Positioning System (GPS) applications.
• Time Interval Error (TIE) modeling is essential for understanding and mitigating clock drift.
• Existing predictive filters may not offer optimal performance in all clock holdover scenarios.

Purpose of the Study:

• To develop an unbiased p-step predictive finite impulse response (FIR) filter for the local clock K-degree time interval error (TIE) polynomial model.
• To derive generic coefficients for a 2-parameter family of polynomial filter gains.
• To generalize the p-step linear (ramp) gain for near-optimal predictive filtering of TIE.

Main Methods:

• Development of a p-step predictive FIR filter based on a K-degree TIE polynomial model.
• Derivation of generic coefficients for a 2-parameter family of polynomial filter gains.
• Generalization of the p-step linear gain for enhanced predictive filtering.

Main Results:

• The proposed filter provides unbiased prediction of TIE.
• Generic coefficients allow for flexible filter gain adjustments.
• The generalized linear gain enables near-optimal predictive filtering of TIE.
• Simulations and real-world GPS-based measurements validate the filter's efficiency in clock holdover.

Conclusions:

• The developed unbiased p-step predictive FIR filter offers significant improvements for GPS-based clock synchronization.
• The filter's efficiency is confirmed in both simulated and practical clock holdover applications.
• The generalized gain provides a robust approach to predictive TIE filtering.