ABOUT THE ACCURACY OF ESTIMATED FREQUENCIES WITH THE PyFEST SOFTWARE

Abstract

This paper presents a method to estimate frequencies with high accuracy that is based on an iterative method. The method consists in applying rectangular windows with different lengths to the acquired or generated signal in order to control the resulted frequency resolution. For each truncated signal we calculate the Discrete Fourier Transform (DFT), finally being overlapped all achieved spectra. So, we obtain a unique spectrum form all DFTs, which have an extremely fine frequency resolution. For the three biggest amplitude values found in the vicinity of the targeted frequency we perform an interpolation and find the regression curve and the maximum of it. This represent the true amplitude of the signal’s targeted harmonic component. The inter-bin position on which this amplitude is found indicates the true frequency. The algorithm is transposed in an application written in the Python programming language and tests are made to find the accuracy of the proposed estimation method. Since we find frequency estimates very close to the generated frequencies, the method and the application are successfully validated.

Keywords

• python programming language
• frequency estimation
• iteration
• pyfest software
• signal processing.