Date of Award


Degree Type


Degree Name

Master of Engineering (ME)


Electrical and Computer Engineering


S.K. Sarna


The problem of estimating the power spectral density of Gastrointestinal (GI) signals is studied. Classical nonparametric methods based on the discrete Fourier Transform are presented along with modern parametric methods which are generally based on autoregressive- moving average (ARMA) time series models. Some algorithms of ARMA parameter estimation are presented under the classification of optimal and suboptimal methods. Two recently proposed ARMA methods are particularly considered and proven to be equivalent. Additive interference between autoregressive signals is shown to produce an ARMA process with equal orders of the MA and the AR parts. This has been made use of in modelling and spectral estimation of the Electrical Control Activity (ECA) in the small intestine.

Four methods for spectral estimation of GI signals have been implemented in a general minicomputer based program.

The performance of the different methods is demonstrated by examples taken from small intestinal ECA. The spatial distribution and the temporal variations have been investigated for the ECA spectra in the small intestine.