Ilmenau Package for Model Order Selection and Evaluation of Model Order Estimation Scheme of Users of MIMO Channel Sounders

João Paulo C. Lustosa da Costa, Florian Roemer, Flávio A. de Castro Jr., Rafael F. Ramos, Lola Sabirova, and Stefanie Schwarz

DOI: 10.14209/sbrt.2011.118

Evento: XXIX Simpósio Brasileiro de Telecomunicações (SBrT2011)

Keywords: Model Analysis. Order Selection Principal Component

Abstract

In several scientific fields, principal component anal- ysis (PCA) is applied to extract information from the data and the first crucial step for PCA is the estimation of the amount of principal components, also known as model order. For instance, for sensor array signal processing, robust techniques for the estimation of the model order are needed, since in practice only a limited number of observations is available. In this paper, we propose a software called Ilmenau package for model order selection (IPM) to compare the several state-of- the-art model order selection schemes in the literature. Since one model order estimation scheme can outperform another scheme for only a certain scenario, it is important to have such a software that evaluates which scheme is the best for a desired scenario. Another major contribution of this paper is to compare the performance of the model order estimation scheme of users of MIMO channel sounders to the state-of-the-art model order selection schemes.

Download