Matlab Rayleigh Fading Channel Simulation?

EDAboard.com Forum Index -> Digital communication -> Matlab Rayleigh Fading Channel Simulation?

Author

Message

davyzhu

Joined: 23 May 2004
Posts: 521
Helped: 3
Location: oriental

06 Mar 2006 10:19 rayleigh fading matlab

Hi all,

I have got a Rayleigh Fading Channel Simulation code by Matlab.

The code list below:
% Rayleigh fading
a = sqrt(0.5)*( randn( 1, symbols_per_frame) + j*randn( 1, symbols_per_frame) );
% complex noise
noise = sqrt(variance)*( randn(1,symbols_per_frame) + j*randn(1,symbols_per_frame) );
% in all
code_fade = a.*code + noise;

Is the code right? If not right, how to modify it?

Best regards,
Davy

manasiri

Joined: 01 Mar 2006
Posts: 234
Helped: 16

07 Mar 2006 12:46 rayleigh channel matlab

Dear sir,

I think that the answer of your problrm is in the following link

Enjoy it.

best regards

http://www.mathworks.com/matlabcentral/fileexchange/loadFileList.do

mehtesham

Joined: 14 Nov 2005
Posts: 111
Helped: 12

12 Mar 2006 22:09 matlab rayleigh channel

your code can be correct.

but in some simulation i did:

% Rayleigh fading

a = randn( 1, symbols_per_frame) + j*randn( 1, symbols_per_frame) ;

% complex noise

noise = randn(1,symbols_per_frame) + j*randn(1,symbols_per_frame) ;

% in all

code_fade = sqrt(SNR/Nt)*a.*code_mod + noise;

Nt is number of Tx antenna

davyzhu

Joined: 23 May 2004
Posts: 521
Helped: 3
Location: oriental

13 Mar 2006 5:07 rayleigh fading channel matlab

Hi mehtesham,

Thank you Smile

I am confused with your formular. If Nt increase, sqrt(SNR/Nt) will decrease?
BTW, code I got is SNR/Nt=0.5. Does it mean normalize?

Best regards,
Davy

Google
AdSense
Google Adsense

13 Mar 2006 5:07 Ads

carpa

Joined: 12 Nov 2004
Posts: 78
Helped: 5

13 Mar 2006 7:32 rayleigh channel matlab code

You need filter the output by a digital filter.

ssstar

Joined: 13 Oct 2004
Posts: 16

13 Mar 2006 15:03 phase ditortion channel+matlab

I am sorry to tell you。
I don't think you are right.
because it is very not accurent, you must know randn() is not very stable.

I suggest that you should look at the papers of Clark,Jakes, and C.xiao.
there are three ways to model Rayleight channel, which are IDFT/WSS/AR.
if you just play , forget it.

davyzhu

Joined: 23 May 2004
Posts: 521
Helped: 3
Location: oriental

13 Mar 2006 15:43 frequency selective channel matlab

Hi,

I have read the Probability book. The phase shift is uniformly distribution. And the fading is Rayleigh distribution.

I am interested in paper about Rayleigh channel.

But what's IDFT/WSS/AR mean?

mehtesham

Joined: 14 Nov 2005
Posts: 111
Helped: 12

14 Mar 2006 9:40 rayleigh channel in matlab

see this paper:

http://web.mit.edu/lizhong/www/publications/tradeoff.pdf

mimomod

Joined: 25 Jan 2006
Posts: 109
Helped: 15

18 Mar 2006 23:48 nakagami distribution matlab

Hi davyzhu,

the matlab codes you wrote are basically right. Off course others could argue that the randn() function is not stable or how to define the right power, etc. But basically, it is right.

But.....when I read your codes I found out that the codes are a little strange for simulating wireless channels. First, the codes generated different fading for each symbol. This happens rarely in practice since the channel will be more or less constant for several symbols, otherwise we have much difficulties to do signal processing at the receiver to get the original data back.

Secondly, the codes generated flat fading channel (more accurately, varying flat fading channel). This fading is the simplest fading in wireless environment. More widely used model is frequency selective fading channel, where you generate a kind of filter with its coefficients are the realizations of flat fading channel as generated by your codes. Better, thus more realistic, you take into account the Doppler effect for fading channel due to movement of TX and/or RX. This generation of fading channel needs more sound knowledge on statistics. You could see the book by Paetzold on fading channels. The book was uploaded in this forum, but sorry that I forget where to find.

best

wee

Joined: 07 Nov 2005
Posts: 29

19 Mar 2006 12:17 simulation rayleigh channel

Hi all,

I am sorry if i got this wrong.

For Multipath fading, can i simply add a tau difference to the orignating signal and then use randn function to create an awgn environment?

(*Sorry for the highjacked*)

Please advice.

Thank you.

Regards.

lincolndsp

Joined: 13 Aug 2005
Posts: 209
Helped: 3
Location: Ukraine

25 Aug 2006 8:09 fading channel simulation

mimomod wrote:

What changes should be made to fix this problem

mimomod wrote:

First, the codes generated different fading for each symbol. This happens rarely in practice since the channel will be more or less constant for several symbols

And is the original code multiplied by Rayleigh fading is a passband signal
or baseband?

Noo10

Joined: 02 Sep 2006
Posts: 7

02 Sep 2006 16:24 rayleigh fading channel simulation

hi all
I want to genarte wireless source code for matlab Ka...

Thanks in Advance

dalysk

Joined: 30 Oct 2006
Posts: 6

29 Jan 2007 14:28 rayleigh fading in matlab

see this site :
http://www.faqs.org/docs/sp/sp-1.html

amihomo

Joined: 09 Jan 2007
Posts: 206
Helped: 23
Location: Tehran,Iran

31 Jan 2007 6:39 simuation of fast fading channel mat lab

davyzhu wrote:

hi,

essentialy the fading process has two effects :
1)a phase distortion which is modeled as a uniform R.V in the interval [-Π,Π]
2)a push distortion which is modeled as a rayleigh , rice , nakagami , ... distribution.

in simulation we assume that the phase distortion in completeley compensated in the receiver and only analyze the effect of push distortion .

based on the above , your code seems to be correct.

regards

gr8_halem

Joined: 31 Jan 2007
Posts: 6

31 Jan 2007 7:05 flat fading channel matlab

may u try this;

chan=rayleighchan;
msx_rx=filter(chan,msg_mod);
msg_noise(z, Smile

=(msg_rx);
end

changfa

Joined: 14 Dec 2005
Posts: 276
Helped: 82

04 Feb 2007 4:49 rayleigh channel in matlab code

davyzhu wrote:

The codes are no problem, but the physical meaning behind them is another story.

As one friend pointed out, your channel changes so fast in time domain that every symbol experiences independent fading. This is rare in reality because system designer want the TX/RX simple and rarely choose such operating point. A simple model is block-fading channel. To model this, you need to know the coherence time of the channel as well as your symbol duration, then you know how many symbols can be assumed experieceing same channel coefficients. For different block, channel is independent.

(For more accurate channel, you need to take Doppler spectrum into account. Doppler spectrum determines how fast your channel is evolved in time. To simulate this, generally a filtering is done. )

waqasbukhari

Joined: 04 Feb 2007
Posts: 5

05 Feb 2007 0:21 matlab code for rayleigh channel

following is the relevant code which i have used in my simulations in one of the projects "design and optimization of cooperative diversity in wireless sensor networks"

%function flatfading

clear;

N =100; %Enter the number of random scaterers

M =1000 %Enter the no of realizations

A = 1 %Enter the value of A

fc =1000; %carrier frequency

Fs=4000; %sampling frequency greater than Nyquist rate

w = 2 * pi * fc ;

t = [0 Sad

1/Fs)

M-1)*(1/Fs)];

L=length(t);

theta = (rand(N,1) * 2 * pi) - pi;

Ak = (rand(N,1) * 2) - 1;

X = zeros(1,L) ;

for m = 1:N

X=X+Ak(m)*cos(w.*t+theta(m));

end

% Analyzing that X(t) is a narrow-band spectral process

Y = fft(X) ;

Y = fftshift(Y) ;

plot(abs(Y)) ;

title( 'Magnitude of fft of a single realisation of X(t)') ;

xlabel( 'Frequency' ) ;

ylabel( 'Magnitude' ) ;

pause

%Calculating XI and XQ

Ak = (rand(N,M) * 2 ) - 1;

theta = (rand(N,M) * 2 * pi) - pi;

XI = zeros(1,M ) ;

XQ = zeros(1,M ) ;

for i=1:N

XI = XI+(Ak(i, Smile

.*cos(theta(i, Smile

));

XQ = XQ+(Ak(i, Smile

.*sin(theta(i, Smile

));

end

%Calculating the central moments and beta values for in and quadrature phase

U2_inphase = moment(XI,2);

U3_inphase = moment(XI,3);

U4_inphase = moment(XI,4);

B1_inphase = (U3_inphase)^2/(U2_inphase)^3

B2_inphase = (U4_inphase)/(U2_inphase)^2

U2_quad = moment (XQ,2);

U3_quad = moment (XQ,3);

U4_quad = moment (XQ,4);

B1_quad = (U3_quad)^2/(U2_quad)^3

B2_quad = (U4_quad)/(U2_quad)^2

%Calculating the envelope r from the M realisations of XI and XQ

r = sqrt(XI .^ 2 + XQ .^ 2);

phi = atan(XQ ./ XI);

R=hist(r,100);

index = 1:100 ;

R_index = (index/length(index)*ceil( max(r)));

stem(R_index,R)

title( 'Histogram of distribution of the Envelope R' ) ;

xlabel('R') ;

ylabel('Frequency of R' ) ;

pause

% including line of sight component

for a = 0:5

XL = a*cos(2*pi*fc*t);

XI_2 = XI + XL;

r = sqrt(XI_2 .^ 2 + XQ .^ 2);

R = hist(r,100);

index = 1:100 ;

R_index = (index/length(index)*ceil( max(r)));

subplot( 2, 3,(a +1));

plot(R_index,R)

end