Adaptive instantaneous frequency ridge extraction based on target tracking for frequency-modulated signals

Area of Science:

• Signal Processing
• Time-Frequency Analysis
• Non-stationary Signal Analysis

Background:

• Instantaneous frequency (IF) is crucial for analyzing non-stationary signals.
• Simultaneously extracting multiple, crossing, and discontinuous IF ridges remains a significant challenge in signal processing.

Purpose of the Study:

• To propose an adaptive method for the simultaneous extraction of multiple, crossing, and discontinuous IF ridges.
• To address the limitations of existing methods in complex IF ridge extraction.

Main Methods:

• The proposed method treats IF ridge values as trackable targets with 'survival', 'death', and 'birth' states.
• Potential peaks in time-frequency representations are identified to mitigate noise.
• An adaptive penalty function guides target state determination and ridge tracking, with slope information used for separating crossing ridges.

Main Results:

• The method demonstrates capability in simultaneously extracting multiple IF ridges.
• It successfully extracts crossing and discontinuous IF ridges.
• Performance validation was conducted using simulated and experimental signals.

Conclusions:

• The developed adaptive IF ridge extraction method effectively addresses the challenge of simultaneous extraction of complex IF ridge structures.
• This technique offers a robust solution for analyzing non-stationary signals with intricate frequency modulations.