Empirical Mode Decomposition in Signal Analysis

Empirical Mode Decomposition (EMD) is an adaptive, data-driven technique for analysing non-stationary and nonlinear signals by decomposing them into a finite set of oscillatory components known as Intrinsic Mode Functions (IMFs). The method relies on an iterative sifting process that identifies local extrema, constructs upper and lower envelopes by interpolation, and extracts components whose envelopes average to zero. Once IMFs are obtained, the Hilbert Transform can be applied to each component to yield instantaneous frequency and amplitude information, forming the basis of the Hilbert–Huang Transform. EMD has proven invaluable across disciplines—from geophysics and mechanical fault diagnosis to biomedical signal interpretation and energy demand forecasting—owing to its ability to adaptively capture time-varying spectral content without requiring a predefined basis. Core challenges include mode mixing, where disparate scales co-exist in a single IMF, and end effects, which introduce boundary distortions during envelope construction. Over the past decade, enhancements such as noise-assisted variants, weighted spline interpolation, revised stop criteria and population-based theoretical models have advanced the robustness, accuracy and computational scaling of EMD. These developments reinforce the global significance of EMD in extracting physically meaningful features from complex real-world data.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Empirical Mode Decomposition in Signal Analysis publication trend

The graph below shows the total number of articles in empirical mode decomposition in signal analysis across all publications each year (not limited to Nature Index journals).

Technical terms

Empirical Mode Decomposition (EMD): an adaptive, data-driven method that decomposes a signal into oscillatory modes via an iterative sifting process without assuming linearity or stationarity.

Intrinsic Mode Function (IMF): a component extracted by EMD that satisfies two conditions: symmetric envelopes defined by local maxima and minima, and a zero local mean, permitting meaningful instantaneous frequency estimation.

Hilbert Transform (HT): a mathematical operator applied to an IMF to derive its analytic signal, from which instantaneous amplitude and frequency can be computed for time-frequency representation.

Mode Mixing: the unintended sharing of a single IMF by multiple signal scales, often addressed through noise-assisted methods or refined interpolation schemes to ensure scale separation.

End Effects: boundary distortions arising during envelope interpolation at the signal edges, mitigated by endpoint extrapolation, local polynomial fitting or modified interpolation algorithms.

References

1. Tutorial on Empirical Mode Decomposition: Basis Decomposition and Frequency Adaptive Graduation in Non-Stationary Time Series. IEEE Access (2023).
2. A Comparative Analysis of Signal Decomposition Techniques for Structural Health Monitoring on an Experimental Benchmark. Sensors (2021).
3. Spectral Analysis of Electricity Demand Using Hilbert–Huang Transform. Sensors (2020).