Graphic STI
logo EPFL
text EPFL
english only
Biomedical Imaging Group
Algorithms
BIG > Algorithms > Continuous-time ARMA Identification
CONTENTS

Home page

News & Events

People

Research

Publications

Tutorials & Reviews

Demos

Download Algorithms

Teaching

Download

arma-estimation.zip

size: 15,072 bytes

date: 19.07.2012

Continuous-time ARMA Identification

Hagai Kirshner, Biomedical Imaging Group, EPFL.

Description

This collection of Matlab functions estimates continuous-time autoregressive moving-average (ARMA) parameters from sampled data. The proposed approach uses exact evaluation of the discrete-domain power-spectrum for calculating the likelihood function of the sampled data. The continuous-time model is assumed to be Gaussian and no constraints are imposed on the sampling interval value. The likelihood function exhibits several local minima and the proposed algorithm aims at finding the global one by using multiple initial conditions.

The files are:

local-minima
Local minima of the likelihood function are obtained by initializing with different initial parameters, as shown here for the case of an AR(2) process with two complex poles

Reference

H. Kirshner, S. Maggio, M. Unser, "A Sampling Theory Approach for Continuous ARMA Identification," IEEE Transactions on Signal Processing, vol. 59, no. 10, pp. 4620-4634, October 2011

Terms of use

The software is freely available for research purposes.

One needs to have our consent before transmitting and/or distributing it further on.

Presented and/or published material that is based on this software should include the relevant citation.

The EPFL makes no warranties of any kind on this software and shall in no event be liable for damages of any kind in connection with the use and exploitation of this technology.