Abstract—This paper explains and demonstrates how to estimate an output disturbance in an auto-regressive model. This method uses the independent component analysis (ICA) technique, which restores source signals from their linear mixtures under the assumption that the source signals are mutually independent. The estimation is achieved by a model whose source signals consist of input and output disturbance, and observed signals consist of input and output. To solve the ICA problem, a natural gradient method based on mutual information is adopted. As a result, in this simulation, the NRR of our proposed method shows an improvement of about 4.0 [dB] compared with that of a conventional method.
Index Terms—Independent Component Analysis (ICA), Blind Signal Separation (BSS), kullback-leibler divergence, Auto Regressive (AR) model.
The authors are with School of Science and Technology, Meiji University, Kawasaki-shi, Kanagawa, 214-8571 Japan. (e-mail: rtanaka@meiji.ac.jp).
Cite: R. Tanaka, K. Kawaguchi, J. Endo, H. Shibasaki, Y. Hikichi, and Y. Ishida, "Estimation of Output Disturbance in Auto-Regressive Model via Independent Component Analysis," International Journal of Modeling and Optimization vol. 3, no. 1, pp. 67-70, 2013.
Copyright © 2008-2024. International Journal of Modeling and Optimization. All rights reserved.
E-mail: ijmo@iacsitp.com