An Efficient and Configurable Preprocessing Algorithm to Improve Stability Analysis.

Area of Science:

• Time Series Analysis
• Metrology
• Aerospace Engineering

Background:

• Allan variance (AVAR) is crucial for assessing experimental time series stability, particularly in space applications like Global Navigation Satellite Systems (GNSS) clock monitoring.
• Space-based time series data often exhibit unique challenges, including outliers, jumps, and missing values, which can compromise accurate clock characterization.
• Robust data preprocessing is essential for reliable stability estimation using AVAR and similar metrics.

Purpose of the Study:

• To develop and implement a robust preprocessing algorithm for experimental time series data.
• To address challenges specific to space clock data, such as nonstationarities and missing values.
• To enhance the accuracy and reliability of Allan variance (AVAR) based stability analysis.

Main Methods:

• A novel preprocessing algorithm was developed to detect and remove anomalous behaviors in time series data.
• The algorithm was implemented in MATLAB, creating robust software for data cleaning.
• The method specifically targets issues like outliers, jumps, and missing data points.

Main Results:

• The proposed algorithm effectively identifies and removes nonstationarities and missing data from experimental time series.
• The preprocessing significantly improves the quality of data used for stability analysis.
• Subsequent Allan variance (AVAR) calculations become more reliable and accurate after data cleaning.

Conclusions:

• The developed preprocessing algorithm provides a reliable solution for analyzing complex time series data from space applications.
• Implementing this method enhances the accuracy of stability estimations derived from Allan variance (AVAR).
• This work contributes to more dependable characterization of clocks in GNSS and other space-based systems.