17-11-2012, 06:01 PM
Nonlinear Active Noise Control with Virtual Sensing Technique
Nonlinear Active Noise Control.pdf (Size: 376.25 KB / Downloads: 29)
Abstract
In this paper a new active noise control algorithm is
proposed which is suitable for nonlinear acoustical paths when
using the virtual sensing technique. Conventional nonlinear ANC
algorithms are suitable only for noise control at the physical
microphone area. The virtual sensing algorithm effectively
controls the noise at a virtual location and is suitable under both
linear and nonlinear primary paths. The algorithm is simulated
under various conditions of nonlinear duct model and proved to
be superior to the commonly employed FXLMS algorithm.
INTRODUCTION
Noise is one of the most widespread health hazards in
industrialised countries today. Human exposure to sustained
or elevated levels of noise has serious detrimental health
effects ranging from nervousness and stress to high blood
pressure and loss of hearing [1]. Reducing the adverse effects
of noise exposure requires a means of minimising unwanted
noise disturbances. It is not possible to mitigate noise at all
places by using passive techniques such as enclosures, barriers.
Active noise control (ANC) [2] is a better technique in that
context particularly for low frequency noise. ANC attenuates
the offending noise by electro-acoustically generating an
“anti-noise”. A local active noise control system minimises
noise disturbances using a secondary sound source (usually a
loudspeaker) to cancel the acoustic pressure measured at a
physical sensor (usually a microphone) [2]. This results in a
small localised zone of quiet being created at the sensor
location, as shown in Fig. 1 (a) [3]. While the noise
disturbance is significantly attenuated at the sensor position,
the surrounding zone of quiet is very small. Additionally, the
sound pressure levels outside the zone of quiet are likely to be
higher than the original disturbance alone.
NONLINEAR ANC ALGORITHM WITH VIRTUAL SENSING
The filter-bank implementation of the FSLMS algorithm is
shown in Fig. 2 [13]. We can think of this adaptive filter as a
bank of linear filters with the inputs modulated by orthogonal
sinusoidal nonlinear functions. In such a scheme, the acoustic
path from the noise source to the physical microphone is a
primary path which can be linear or nonlinear.
SIMULATION RESULTS AND DISCUSSION
The effectiveness of the proposed nonlinear ANC algorithm
with virtual sensing is verified in simulated experiments of an
acoustic duct. A reduced modal model of a rectangular
acoustic duct given in [8] is used with the first six acoustic
modes. As shown in Fig. 3, the position of the secondary loud
speaker, the physical microphone and the virtual microphone
are fixed at a distance of 3m, 3.5m and 4m respectively within
a rectangular duct model of length 4.2m.