You are on page 1of 12

Presentation on

Application of Adaptive
Filters
Prepared by - Urja Shah
(Me 1 CSE)
Guided by- Usha Mam
Abhay Sir

Adaptive filters
Adaptive filters modify their characteristics to achieve
certain objectives by automatically updating their
coefficients
An adaptive filter consists of two distinct parts a
digital filter to perform the desired filtering, and an
adaptive algorithm to adjust the coefficients (or
weights) of the filter.

Some applications
Adaptive antenna systems
Digital communication receivers
Adaptive noise cancelling techniques
System modelling
Adaptive channel equalization

Adaptive noise cancellation


Cellular phones in high noise environments
Corrupts speech and degrades performance
Noise canceller employs adaptive filter with LMS
algorithm to cancel
To apply the reference noise picked up by the reference
sensor must be highly correlated with the noise
components in the primary signal
Requires a close spacing.

The crosstalk effect


The signal components from the signal source is picked
up by the reference sensor.
It cancels the desired signal along with the undesired
noise.\

Elimination of crosstalk:- FIRST


TECHNIQUE
Placing the primary sensor far away from the reference
sensor.
LIMITATION OF THIS METHOD
This arrangement requires a large-order filter
Not always feasible to place the reference sensor far
away from the signal source

SECOND TECHNIQUE
To place an acoustic barrier between primary and
reference sensor
LIMITATIONS: Many applications do not allow an acoustic barrier
between sensors,
Barrier may reduce the correlation of the noise
component in the primary and reference signals.

THIRD TECHNIQUE
To control the adaptive algorithm to update filter coefficients only
during the silent intervals in the speech.
LIMITATIONS: Depends on a reliable speech activity detector that is very
application dependent
Fails to track the environment changes during the speech periods.
Microphone array techniques are used to improve performance of
noise cancellation.

AN EXAMPLE: As shown in Figure assume s(n) is a sinewave, x(n) is a


white noise and P(z) is a simple FIR system. We use the
adaptive FIR filter with the LMS algorithm for noise
cancelation.
The adaptive filter will approximate P(z), and thus its
output y(n) will converge to x(n) in order to cancel it.
Therefore, the error signal e(n) will gradually approach
the desired sinewave s(n), as shown in fig

THANK YOU

You might also like