Delay Estimation Using Instantaneous Frequency and Phase Difference--Simulation Study.

Area of Science:

• Signal Processing
• Electrical Engineering
• Physics

Background:

• Existing phase-based delay estimators often smooth frequency estimates, potentially losing valuable information from signal interferences.
• Traditional methods may misinterpret phase differences, leading to errors when instantaneous frequency varies significantly.

Purpose of the Study:

• To propose a novel time-domain delay estimation method utilizing the slope of a best fit line in the instantaneous frequency-phase difference plane.
• To address limitations of existing phase-based estimators by incorporating instantaneous frequency at all sample points and redefining error metrics.

Main Methods:

• Calculating instantaneous frequency at individual sample points, including interference spikes.
• Defining estimation error as deviation from a best fit line, not just an averaged phase difference.
• Deriving a Gauss-Markov least-squares best fit line and developing an iterative variant to improve accuracy and reduce aliasing.

Main Results:

• The proposed iterative phase least squares (PLS) method outperforms correlation phase Loupas at high signal-to-noise ratios (SNRs).
• Performance gains of 30-50 dB in root-mean-square error (RMSE) are observed compared to correlation phase Loupas under optimal conditions.
• The iterative approach effectively mitigates aliasing for larger delays and enhances overall estimation accuracy.

Conclusions:

• The proposed time-domain delay estimation technique offers superior performance, particularly at high SNRs, by leveraging instantaneous frequency information.
• Iterative PLS provides a more robust and accurate method for delay estimation compared to conventional approaches like correlation phase Loupas.
• The study highlights the importance of considering instantaneous frequency variations and redefining error metrics for improved phase-based estimation.