Vous êtes sur la page 1sur 5

13

Experiment 5:
Aim: To pass a BPSK signal through a frequency selective fading channel. Perform
channel equalization using LMS adaptive filter. And observe BER vs. SNR plot for
various SNRs.
Theory:
Multipath Fading: If the channel possesses a constant gain and linear phase response over a
bandwidth that is smaller than the bandwidth of the transmitted signal, then the channel
creates frequency selective fading on the received signal. Under such condition the channel
impulse response has a multipath delay spread which is greater than the reciprocal bandwidth
of the transmitted message waveform. When this occurs, the received signal includes multiple
versions of the transmitted waveform which are attenuated (faded) and delayed in time, and
hence the received signal is distorted in time. Frequency selective fading is due to the time
dispersion of the transmitted symbol within the channel. Thus the channel includes inter-
symbol interference (ISI).
Adaptive Channel Equalization: Intersymbol interference (ISI) caused by multipath
bandlimited (frequency selective) time dispersive channel distorts the transmitted signal
causing bit error at the receiver. ISI has been recognized as a major drawback to high speed
data transmission. Equalization is a technique used to combat the ISI.

An equalizer is usually implemented at baseband or at IF in a receiver. Since the baseband
complex envelope expression can be used to represent bandpass waveforms, the channel
response, demodulated signal and adaptive equalizer algorithms are usually simulated and
implemented at baseband.
If x(t) be the original information signal, if f(t) be the combined complex baseband impulse
response of the transmitter, channel and the RF/IF sections of the receiver, the signal received
by the equalizer can be expressed as:
y(t) =x(t)

(t) +n
b
(t)
Where, f*(t) is the complex conjugate of f(t) and n
b
(t) is the baseband noise at the input of the
equalizer. In modern communication systems, as equalizer is implemented in digital
computers, the input to the equalizer in discrete time can be expressed as
y(n) =x(n)

(n) +n
b
(n)
Adaptive equalizers are used to estimate x(n) of x(n) from y(n) at the receiver. An equalizer
is basically an FIR filter that changes its parameters to match the desired output. One most
commonly used equalizer is the Least Mean Square (LMS) equalizer. The update algorithm
of LMS equalizer is given by:
w
n+1
=w
n
+pc(n)x
n


14

Where, w
n
is the present filter parameter vector, w
n+1
is the next filter parameter vector, x
n

is
the complex conjugate of the input vector {x(n), x(n 1), x(n 2), , x(n L + 1)}, p is the
step size of the adaptive equalizer and c(n) =x(n) x(n) is the error.

Figure: A basic linear equalizer during training

It is to be noted that the value of actual signal x(n) must be known to the signal to calculate
the error e(n) in order to update the adaptive equalizer. However, practically the receiver
cannot have the values of x(n). Therefore, some pilot symbols are transmitted along with the
actual payload to train the adaptive equalizer. These pilots are inserted either at the
beginning, or at the end or at the middle of a message block. These pilots are known to the
receiver. These are used to train the adaptive filter. Once the filter is trained, the resultant FIR
filter acts as an inverse of the complex filter f(n). Thus we can have:
x(n) =y(n) w(n)

Program
%mul t i pat h channel equal i zat i on ( bl ock t ype channel )
cl ear al l ;
cl ose al l ;
cl c
%def i ni ng modul at or and demodul at or
M=2;
hm=modem. pskmod( M) ;
hd=modem. pskdemod( M) ;
%Lengt h i ni t i al i zat i on
15

N=10000;
Nf =1000;
Np=200;
r ng( ' shuf f l e' ) ;
%Cr eat i ng pi ol t
pi l ot _bi t s=r andi ( [ 0, 1] , 1, Np) ;
pi l ot _symb=modul at e( hm, pi l ot _bi t s) ;
ch_or d=4;
nf _msg=Nf - Np;
N1=cei l ( N/ nf _msg) ;
r chn1=1/ sqr t ( 2) *( r andn( ch_or d, N1) +1j *r andn( ch_or d, N1) ) ;
r chn1=r chn1/ max( max( r chn1) ) ;
f or k=1: N1
r chn1( : , k) =sor t ( r chn1( : , k) , ' descend' ) ;
end
%r chn1=r chn1/ nor m( r chn1) ;
r chn=ones( Nf , ch_or d) ;
f or k=1: Nf : N
f or k1=1: ch_or d
r chn( k: k+Nf - 1, k1) =r chn1( k1, ( k- 1) / Nf +1) *ones( Nf , 1) ;
end
end
f i gur e,
subpl ot 211
pl ot ( abs( r chn) )
t i t l e( ' Channel Magni t ude' )
subpl ot 212
pl ot ( angl e( r chn) / pi )
t i t l e( ' Channel Phase/ pi ' )
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%I nput si gnal
x=r andi ( [ 0, 1] , 1, N) ;
i f ( mod( N, nf _msg) ~=0)
N_add=cei l ( N/ nf _msg) *nf _msg- N;
N=N+N_add;
x=[ x, zer os( 1, N_add) ] ;
end
snr =- 10: 10;
N_f ul l =Nf *N/ nf _msg;
Tx_chan=zer os( 1, N_f ul l ) ;
count =1;
f or k=1: nf _msg: N
x_t emp=x( k: k+nf _msg- 1) ;
mod_msg=modul at e( hm, x_t emp) ;
Tx_symb=[ mod_msg( 1: nf _msg/ 2) , pi l ot _symb, mod_msg( nf _msg/ 2+1: nf _msg) ] ;
w=r chn1( 1: ch_or d, ( ( count - 1) / Nf +1) ) ;
chan_out =f i l t er ( w, 1, Tx_symb) ;
Tx_chan( count : count +Nf - 1) =chan_out ;
count =count +Nf ;
end
ber _ex=zer os( si ze( snr ) ) ;
ber _t h=zer os( si ze( snr ) ) ;
eq_or d=5;
mu=0. 001;
f or k=1: l engt h( snr )
Tx_si g=awgn( Tx_chan, snr ( k) ) ;
Rx=zer os( 1, N) ;
count =1;
f or k1=1: Nf : N_f ul l
Tx_t emp=Tx_si g( [ ( k1: k1+nf _msg/ 2- 1) , ( k1+nf _msg/ 2+Np: k1+Nf - 1) ] ) ;
pi l ot _r x=Tx_si g( k1+nf _msg/ 2: k1+nf _msg/ 2+Np- 1) ;
16

%Adapt i ve equal i zat i on ( MSE al gor i t hm)
h=zer os( 1, eq_or d) ;
er 1=zer os( Np- eq_or d+1, 1) ;
f or k2=eq_or d: Np
pk=pi l ot _r x( k2: - 1: k2- eq_or d+1) ;
s=pk*conj ( h' ) ;
er 1( k2) =pi l ot _symb( k2) - s;
h=h+mu*conj ( pk) *( er 1( k2) ) ;
end
Rx_t emp=f i l t er ( h, 1, Tx_t emp) ;
Rx( count : count +nf _msg- 1) =demodul at e( hd, Rx_t emp) ;
count =count +nf _msg;
end
ber _t emp=l engt h( f i nd( x- Rx) ~=0) ;
ber _ex( k) =ber _t emp/ l engt h( Rx) ;
ber _t h( k) =ber f adi ng( snr ( k) , ' psk' , M, 1) ;
% f i gur e, semi l ogy( abs( er 1) ) ;
end
f i gur e
l 1=semi l ogy( snr , ber _ex, ' b- s' ) ;
hol d on
l 2=semi l ogy( snr , ber _t h, ' r ' ) ;
l egend( ' BER Exper i ment al ' , ' BER Theor et i cal ' )
xl abel ( ' SNR ( dB) ' ) ;
yl abel ( ' BER' ) ;


Channel Response:


0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
0
0.5
1
Channel Magnitude
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
-1
-0.5
0
0.5
1
Channel Phase/pi
17

Output:



-10 -8 -6 -4 -2 0 2 4 6 8 10
10
-2
10
-1
10
0
SNR (dB)
B
E
R


BER Experimental
BER Theoretical

Vous aimerez peut-être aussi