Frequency Estimation using Recursive Least Square Technique

Abstract

When the frequency is known, adaptive Fourier analyzers determine the coefficients of the sine and cosine terms of a noisy sinusoidal signal. However, in actual use, the frequencies could be different from their predicted values. This anomaly is known as a frequency mismatch (FM). The performance of the Recursive Least Square (RLS) Fourier analyzer while the FM is available was investigated in this work. The amount of computations is only marginally increased due to the comparatively quick convergence to steady state of this RLS- based method. Based on mean square error (MSE) and mean square deviation (MSD), simulations are run to demonstrate the superior performance of RLS over the Least Mean Square (LMS) approach.