Algoritmo para la Estimación del Periodo de una Señal Muestreada

Abstract

If there are only samples of a signal generally it is not possible to determine the period, but only an approximation is achieved. This article describes an algorithm to approximate that period, based on the Discrete Fourier Transform is proposed. If this transform is evaluated multiples of the actual period, it presents some harmonics with zero amplitude. Thus, the estimated period is one that minimizes the value of these harmonics. To validate the algorithm functions are used with and without noise, which is that the relative error is always equal to or less than a quarter of the time between two samples. As an example of use of the algorithm, the actual period is estimated physiological signals, and frequency of the grid is also estimated, in real time.