Search-Enhanced Instantaneous Frequency Detection Algorithm: A Preliminary Design

TL;DR

This paper introduces SEIFD, a novel algorithm that employs genetic algorithms to accurately detect sinusoidal components in complex, nonlinear, non-stationary time-series data, enhancing frequency analysis.

Contribution

The paper proposes a new search-enhanced method using genetic algorithms for instantaneous frequency detection in challenging time-series data.

Findings

Successfully identifies start-time, end-time, frequency, and phase of sinusoidal components.

Demonstrates effectiveness in nonlinear, non-stationary data environments.

Provides a preliminary design for the SEIFD algorithm.

Abstract

This paper presents a method developed for finding sinusoidal components within a nonlinear non-stationary time-series data using Genetic Algorithm (GA) (a global optimization technique). It is called Search-Enhanced Instantaneous Frequency Detection (SEIFD) algorithm. The GA adaptively define the configuration of the components by simulating the solution finding process as a series of genetic evolutions. The start-time, end-time, frequency, and phase of each of these components are identified once convergence in the implementation is achieved.