Main Content

High Resolution Spectral Analysis in Simulink

This example shows how to perform high resolution spectral analysis in Simulink® using the Spectrum Analyzer block and the Spectrum Estimator block.

For comparison purposes, a traditional averaged modified periodogram (Welch's) method is also shown. For a similar example in MATLAB®, see High Resolution Spectral Analysis in MATLAB.

Using Spectrum Analyzer

The SpectrumAnalyzerFilterBank model illustrates the high resolution capabilities of filter bank-based spectral estimation compared to the Welch's method. The filter bank-based spectral estimation has lower noise floor.

Consider the following case. Three sinusoids at 170 kHz, 200 kHz and 205kHz with the amplitudes [1e-5 1 2]. The first sinusoid is completely missed by the rectangular window estimate. The filter bank estimate provides better resolution and better isolation of three tones.

Open and simulate the SpectrumAnalyzerFilterBank model.

Close the model.

Using Spectrum Estimator

Numerical computations for high resolution spectral estimation shown above can also be modeled in Simulink using the Spectrum Estimator block. The SpectrumEstimatorFilterBank model illustrates the high resolution capabilities of filter bank-based spectrum estimation and lower noise floor compared to the Welch's method, using Simulink. Array plot is used to visualize the results. Array plot provides a convenient way of plotting the spectrum estimates. Values are shown in dBm, but Watts or dBW could easily be used instead.

Open and simulate the SpectrumEstimatorFilterBank model.

Close the model.

Related Topics

External Websites