A regression approach for simulating feedforward active noise control

TR Number
Date
1995-05-01
Journal Title
Journal ISSN
Volume Title
Publisher
Acoustical Society of America
Abstract

Regression analysis is used to examine feedforward active noise control from a statistical point of view. Since numerical techniques for simulating feedforward active noise control in the frequency domain are mathematically similar to linear least-squares regression, two regression-based numerical methods can be applied to control problems. The first uses regression diagnostics such as the F-test, the t-test, and confidence intervals to model the effects of error sensor measurement noise. The second uses collinearity diagnostics to address a form of numerical ill conditioning that can corrupt the results. The regression diagnostics allow realistic modeling of random measurement error; the collinearity diagnostics help avoid numerical difficulties that might otherwise go undetected. Numerical results are given for a structural-acoustic control problem involving a fluid-loaded cylindrical shell. 1995 Acoustical Society of America

Description
Keywords
Active noise control, Acoustic noise measurement, Numerical modeling, Statistical analysis
Citation
Ruckman, C. E., & Fuller, C. R. (1995). A regression approach for simulating feedforward active noise control. Journal of the Acoustical Society of America, 97(5), 2906-2918. doi: 10.1121/1.411857