04-06-2012, 11:02 AM
DETECTION GUIDED LMS BASED CHANNEL EQUALIZATION
[attachment=24077]
Abstract
The main objective of this paper is to suppress echoes
arising from non-line-of-sight (NLOS) components in
wireless communication systems. In wireless
communications, the channel has a typically long impulse
response and the signal is highly correlated speech. The
Standard Least Mean Square (LMS) channel estimator
is used as a basis to this paper. In order to reduce the
unfavorable effect of ‘channel dimension’ on the
performance of the LMS estimator, a detection scheme is
incorporated. This scheme allows discrimination between
active and inactive taps of the unknown channel.
However, the detection scheme fails under colored input
signal conditions. Hence, the algorithm is modified to
include a Tap Decoupling feature, which reverses the
effect of the colored input signal. We proposed methods
which are structurally consistent for both White and
colored input signals. We investigated the principles of
LMS adaptive Equalizer which uses the LMS channel
estimator to estimate the NLOS components and
subsequently suppress these components. An Adaptive
Equalizer which incorporates Tap Decoupling with
Detection scheme is designed and implemented. All the
Mat Lab simulation results demonstrates that the design
approach investigated in this project is a promising
alternative for suppressing echoes due to NLOS
Components in the Wireless communication System.
Keywords- LMS Channel Estimator, Non- line of sight
components (NLOS),Tap Decoupling, Eqalizer,Detection
Schemes, filter Taps, Adaptive Filters active Taps,
Inactive Taps, Sqared Euclidian Norm.
Introduction
In this paper we present the issues and problems that are
involved in various methods which are proposed. The
LMS Adaptive FIR estimator applied via configuration
shown in Figure 1 has wide range of applications
particularly in Signal Processing, one such a application
is echo cancellation. In this paper we consider LMS
estimation of a channel which may be well approximated
by an FIR model with active or nonzero taps. An example
of such a channels includes room acoustic echo paths,
echo paths in telephone network and mobile radio
channels. From the previous analysis it is evident that the
LMS estimation of channels should be greatly enhanced
if and only if the active taps are estimated. This leads to
the proposal of a number of LMS procedures
incorporating active tap detection methods.
The detection guided LMS Estimation approach
requires an activity measure and activity threshold. To
determine the tap to be active, the value of activity
measure must be above a minimum value called activity
threshold. The activity measure and activity threshold
proposed here are structurally consistent for white input
signals. The algorithms proposed in this paper are not
based on the structural consistency and they provide
asymptotic performance to the white input conditions.
In Wireless communication channel, the channel has
a typically long impulse response and the signal is highly
correlated speech. The correlation within the input signal
causes coupling amongst the outputs oh the unknown
channel taps. Due to this some inactive taps may appear
as active taps leads to increasing in activity noise level.
Therefore such a effects leads to failure of the algorithms-
[1] and [2] proposed in this paper. This algorithms [1]
and [2] provides good results in the case of white input
signal but failed under the colored input signals because
colored input signals are highly correlated signals. Hence
this algorithm is modified so that which reserves the
effects of colored input signal by modifying the activity
threshold. This modified algorithm is [3].
The most popular application of adaptive filters
named Inverse Modeling, also known as Deconvolution
has found extensive use in various engineering
disciplines. The most popular application of Inverse
Modeling is in communications where an inverse model,
which is also called as an Equalizer, is used to reduce the
channel distortion. Hence we proposed a new system
shown in figure 2, which is an LMS Estimator in series
with Equalizer. The most important effect occurred in
communication channel is Pulse spreading effect is due to
nonzero response of the channel causes Inter Symbol
Interference (ISI). This ISI can be eliminated using
Equalization of a channel using an Equalizer. The
Equalizer proposed in this paper is constructed with
Adaptive filter. This equalizer increases the asymptotic
performance of an algorithms proposed in [1],[2] and [3]
also decreases asymptotic errors than the errors estimated
by the proposed system shown in figure 2.
[attachment=24077]
Abstract
The main objective of this paper is to suppress echoes
arising from non-line-of-sight (NLOS) components in
wireless communication systems. In wireless
communications, the channel has a typically long impulse
response and the signal is highly correlated speech. The
Standard Least Mean Square (LMS) channel estimator
is used as a basis to this paper. In order to reduce the
unfavorable effect of ‘channel dimension’ on the
performance of the LMS estimator, a detection scheme is
incorporated. This scheme allows discrimination between
active and inactive taps of the unknown channel.
However, the detection scheme fails under colored input
signal conditions. Hence, the algorithm is modified to
include a Tap Decoupling feature, which reverses the
effect of the colored input signal. We proposed methods
which are structurally consistent for both White and
colored input signals. We investigated the principles of
LMS adaptive Equalizer which uses the LMS channel
estimator to estimate the NLOS components and
subsequently suppress these components. An Adaptive
Equalizer which incorporates Tap Decoupling with
Detection scheme is designed and implemented. All the
Mat Lab simulation results demonstrates that the design
approach investigated in this project is a promising
alternative for suppressing echoes due to NLOS
Components in the Wireless communication System.
Keywords- LMS Channel Estimator, Non- line of sight
components (NLOS),Tap Decoupling, Eqalizer,Detection
Schemes, filter Taps, Adaptive Filters active Taps,
Inactive Taps, Sqared Euclidian Norm.
Introduction
In this paper we present the issues and problems that are
involved in various methods which are proposed. The
LMS Adaptive FIR estimator applied via configuration
shown in Figure 1 has wide range of applications
particularly in Signal Processing, one such a application
is echo cancellation. In this paper we consider LMS
estimation of a channel which may be well approximated
by an FIR model with active or nonzero taps. An example
of such a channels includes room acoustic echo paths,
echo paths in telephone network and mobile radio
channels. From the previous analysis it is evident that the
LMS estimation of channels should be greatly enhanced
if and only if the active taps are estimated. This leads to
the proposal of a number of LMS procedures
incorporating active tap detection methods.
The detection guided LMS Estimation approach
requires an activity measure and activity threshold. To
determine the tap to be active, the value of activity
measure must be above a minimum value called activity
threshold. The activity measure and activity threshold
proposed here are structurally consistent for white input
signals. The algorithms proposed in this paper are not
based on the structural consistency and they provide
asymptotic performance to the white input conditions.
In Wireless communication channel, the channel has
a typically long impulse response and the signal is highly
correlated speech. The correlation within the input signal
causes coupling amongst the outputs oh the unknown
channel taps. Due to this some inactive taps may appear
as active taps leads to increasing in activity noise level.
Therefore such a effects leads to failure of the algorithms-
[1] and [2] proposed in this paper. This algorithms [1]
and [2] provides good results in the case of white input
signal but failed under the colored input signals because
colored input signals are highly correlated signals. Hence
this algorithm is modified so that which reserves the
effects of colored input signal by modifying the activity
threshold. This modified algorithm is [3].
The most popular application of adaptive filters
named Inverse Modeling, also known as Deconvolution
has found extensive use in various engineering
disciplines. The most popular application of Inverse
Modeling is in communications where an inverse model,
which is also called as an Equalizer, is used to reduce the
channel distortion. Hence we proposed a new system
shown in figure 2, which is an LMS Estimator in series
with Equalizer. The most important effect occurred in
communication channel is Pulse spreading effect is due to
nonzero response of the channel causes Inter Symbol
Interference (ISI). This ISI can be eliminated using
Equalization of a channel using an Equalizer. The
Equalizer proposed in this paper is constructed with
Adaptive filter. This equalizer increases the asymptotic
performance of an algorithms proposed in [1],[2] and [3]
also decreases asymptotic errors than the errors estimated
by the proposed system shown in figure 2.