Error of the signal cyclic component period estimation using the AR model

This paper deals with a method for the estimation of the signal cyclic component´s period using an autoregressive model. The properties of the method are explored and quantified in the case of its application to short sample size signals. A number of computer experiments has been performed using the harmonic signal corrupted by noise and the autoregressive model of the second order. The results of the experiments are represented in the graphical form. It could be noticed that the mean value of the errors could result to significant values while the variance of the error is generally almost negligible.