Prev: Custom Colormap
Next: Plot points on map as varying size dots or circles.
From: Kashif on 23 Jun 2010 11:43
Hi All,

How shall I Generate Rayleigh Channel (h) with a variance other than 1 using randn.
Kinldy also tell how can I verify it.

The problem I am facing is

N=10000;

********************************************************
h=rand(N,1)+j*randn (N,1)/sqrt(2)
% gives var(h)=~0.5 and mean (real(h))=~0.5)
********************************************************
h= sqrt(0.5)*rand(N,1) + j*sqrt(0.5)*randn (N,1)
% gives var(h)=~0.5 and mean (real(h))=~0.35)
********************************************************
h=rand(N,1)+j*randn (N,1)
% gives var(h)=~1 and mean (real(h))=~0.5)
********************************************************
if Z=X+Y, then
As per var(Z)=var(X)+var(Y). Why this formula is not working for above.

Secondly why mean is not zero.
From: Roger Stafford on 23 Jun 2010 16:08
"Kashif " <rajakashif(a)gmail.com> wrote in message <hvta2a$7g8$1(a)fred.mathworks.com>...
> Hi All,
>
> How shall I Generate Rayleigh Channel (h) with a variance other than 1 using randn.
> Kinldy also tell how can I verify it.
>
> The problem I am facing is
>
> N=10000;
>
> ********************************************************
> h=rand(N,1)+j*randn (N,1)/sqrt(2)
> % gives var(h)=~0.5 and mean (real(h))=~0.5)
> ********************************************************
> h= sqrt(0.5)*rand(N,1) + j*sqrt(0.5)*randn (N,1)
> % gives var(h)=~0.5 and mean (real(h))=~0.35)
> ********************************************************
> h=rand(N,1)+j*randn (N,1)
> % gives var(h)=~1 and mean (real(h))=~0.5)
> ********************************************************
> if Z=X+Y, then
> As per var(Z)=var(X)+var(Y). Why this formula is not working for above.
>
> Secondly why mean is not zero.
- - - - - - -
Try

r = sigma*(randn(N,1);
h = r.*exp(j*2*pi*rand(N,1));

where sigma is the square root of the desired variance.

Roger Stafford
 | 
Pages: 1
Prev: Custom Colormap
Next: Plot points on map as varying size dots or circles.