Far-Field Speech Recognition Using Multivariate Autoregressive Models

This paper considers the prediction problems of a k-dimensional, pth order autoregressive process with unstable but non-explosive roots and dependent error variables. The estimated predictor has been shown to be asymptotically equivalent to the optimal predictor. An expression for the meansquare error of the estimated predictor has also been derived .

Download Full-text

Multivariate autoregressive models for classification of spontaneous electroencephalographic signals during mental tasks

IEEE Transactions on Biomedical Engineering ◽

10.1109/10.661153 ◽

1998 ◽

Vol 45 (3) ◽

pp. 277-286 ◽

Cited By ~ 255

Author(s):

C.W. Anderson ◽

E.A. Stolz ◽

S. Shamsunder

Keyword(s):

Autoregressive Models ◽

Multivariate Autoregressive Models ◽

Electroencephalographic Signals ◽

Mental Tasks ◽

Multivariate Autoregressive

Download Full-text

Multivariate autoregressive models

Statistical Parametric Mapping ◽

10.1016/b978-012372560-8/50040-1 ◽

2007 ◽

pp. 534-540 ◽

Cited By ~ 6

Author(s):

W. Penny ◽

L. Harrison

Keyword(s):

Autoregressive Models ◽

Multivariate Autoregressive Models ◽

Multivariate Autoregressive

Download Full-text

Using the Eigenvalues of Multivariate Spectral Matrices to Achieve Cutting Direction and Sensor Orientation Independence

Journal of Manufacturing Science and Engineering ◽

10.1115/1.2123067 ◽

2005 ◽

Vol 128 (1) ◽

pp. 350-354 ◽

Cited By ~ 7

Author(s):

John T. Roth

Keyword(s):

Cutting Tools ◽

Autoregressive Models ◽

Triaxial Accelerometer ◽

Industrial Environment ◽

Monitoring Techniques ◽

Monitoring Technique ◽

Sensor Orientation ◽

Multivariate Autoregressive Models ◽

Cutting Direction ◽

Multivariate Autoregressive

There is a strong need for monitoring techniques capable of tracking the health of cutting tools under varying conditions. Unfortunately, most monitoring techniques are dependent on the cutting direction and/or the sensor orientation, limiting their effectiveness in the typical industrial environment. With this in mind, this research develops a monitoring technique that is independent of both of these factors. This is accomplished by using multivariate autoregressive models that are fit to the output from a triaxial accelerometer. The work shows that the eigenvalues of multivariate spectral matrices, calculated at the machining frequencies, are not only sensitive to the condition of the tool but are also independent of the direction of cutting and the orientation of the sensor. This independence is verified experimentally through tests conducted under a variety of cutting directions and sensor orientations.

Download Full-text