Particle filter for multipath time delay tracking from correlation functions in deep water.

Area of Science:

• Signal Processing
• Acoustics
• Data Analysis

Background:

• Multipath propagation causes signal distortion, complicating time delay estimation.
• Conventional methods often rely on correlation pulses and prior noise knowledge, limiting their effectiveness.

Purpose of the Study:

• To develop a particle filtering-based approach for tracking multipath time delays.
• To overcome limitations of conventional methods by tracking signal peaks instead of correlation pulses.

Main Methods:

• Utilized particle filtering to exploit time-evolving correlations in background noise.
• Introduced local peaks in the measurement equation, removing assumptions on reference signals and noise covariance.
• Extracted time-evolving marginal posterior probability densities to quantify time delay uncertainty.

Main Results:

• Successfully tracked multipath time delays using simulated and experimental deep-water acoustic data.
• Demonstrated significant advantages over conventional state-space, multiple hypothesis tracking, and peak amplitude detection methods.

Conclusions:

• The proposed particle filtering approach effectively tracks multipath time delays.
• This method provides a robust and accurate alternative for underwater acoustic signal analysis.