Fading

Hello All

My ques might sound simple but I have been facing some trouble
understanding certain concepts. I will write here my understandings and I
would be greatful if you can tell if I am right ot wrong. If I am wrong
please help me understand what is it actually.

1. Frequency Selective fading:
 when a signal arrives via multiple path. This can be modelled as below:
 Lets say three path channel h=[h1 h2 h3]; So the o/p of channel is 
 y=conv(h,x); where x is transmitted signal

2. Flat rayleigh fading:
   Signal arrives via one path and can be modelled as below:
   h=sqrt(0.5)*(randn(1,1000)+j*randn(1,1000));

3. Then what is rayleigh frequency selective fading?
  Does that mean h is rayleigh random variable and can be generated as
  h=randn(1,3)+j*randn(1,3); And the o/p of the channel is
  y=conv(h,x);

I would really appreciate your guidance.

Thanks

Chintan

cpshah99 wrote:
> Hello All
> 
> My ques might sound simple but I have been facing some trouble
> understanding certain concepts. I will write here my understandings and I
> would be greatful if you can tell if I am right ot wrong. If I am wrong
> please help me understand what is it actually.
> 
> 1. Frequency Selective fading:
>  when a signal arrives via multiple path. This can be modelled as below:
>  Lets say three path channel h=[h1 h2 h3]; So the o/p of channel is 
>  y=conv(h,x); where x is transmitted signal

You have described a possible cause and confused it with the phenomenon 
itself. Differential reflection in the ionosphere and troposphere is 
another cause. Frequency-selective fading involves some frequencies in a 
signal fading to a greater extent than others.

> 2. Flat rayleigh fading:
>    Signal arrives via one path and can be modelled as below:
>    h=sqrt(0.5)*(randn(1,1000)+j*randn(1,1000));

Fading can be (sort of) thought of as low-frequency noise or 
supplementary modulation. What limits the frequencies in 'h' above?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

cpshah99 <cpshah99@rediffmail.com> wrote:

>1. Frequency Selective fading:
> when a signal arrives via multiple path. This can be modelled as below:
> Lets say three path channel h=[h1 h2 h3]; So the o/p of channel is 
> y=conv(h,x); where x is transmitted signal
>
>2. Flat rayleigh fading:
>   Signal arrives via one path and can be modelled as below:
>   h=sqrt(0.5)*(randn(1,1000)+j*randn(1,1000));
>
>3. Then what is rayleigh frequency selective fading?
>  Does that mean h is rayleigh random variable and can be generated as
>  h=randn(1,3)+j*randn(1,3); And the o/p of the channel is
>  y=conv(h,x);

Usually, Rayleigh fading means that the envelope of the channel
impulse response follows a Rayleigh function.  A Rayleigh function
is zero at t=0, reaches a peak at some non-zero t, then declines
to zero at t=infinity.  Within this envelope, the number,
density, and pattern of individual impulses are further parameters
of your channel model.  A Rayleigh channel might naturally arise
from certain ensembles of large numbers of reflections, but
this is not the only possible type of Rayleigh channel.

Flat fading vs. frequency-selective fading has to do with the
variance of the magnitude of the channel freqeuncy response,
over your comminucation system's bandwidth.  The same channel
may be considered flat on a narrowband system but frequency-selective
on a wideband system.  Also, both Rayleigh channels and non-Rayleigh
channels might be either flat or frequency selective.

Hope this helps.

Steve

On Jan 23, 6:05�pm, spop...@speedymail.org (Steve Pope) wrote:

> Usually, Rayleigh fading means that the envelope of the channel
> impulse response follows a Rayleigh function. �A Rayleigh function
> is zero at t=0, reaches a peak at some non-zero t, then declines
> to zero at t=infinity. �Within this envelope, the number,
> density, and pattern of individual impulses are further parameters
> of your channel model. �A Rayleigh channel might naturally arise
> from certain ensembles of large numbers of reflections, but
> this is not the only possible type of Rayleigh channel.
>


This is a very interesting meaning for Rayleigh fading
that I have not encountered before. A reference where
more details can be found would be much appreciated.

Hello All

Thanks a lot for your replies.

But I am still confused from the simulation point of view.

After your replies, I am more confused. 

I know that rayleigh and non-rayleigh channels can be flat as well as
frequency selective. Also doppler effect plays an important role for
fading.

actually what i am trying is to model underwater channel. Now, in
underwater even if the transmitter and receiver are not moving, there will
be time variations, right?. And from the previous experiments, it was seen
that there exist a direct path, means it follows rician model.

Now, my problem is how simulate this rician model. Because rician is
direct path + rayleigh fading.

I would really appreciate your guidance and if there is any good tutorial
which clearly explains how to simulate this particular thing, that will be
great.

Thanks again for your time and help.

Chintan

dvsarwate@yahoo.com schrieb:
> On Jan 23, 6:05 pm, spop...@speedymail.org (Steve Pope) wrote:
> 
>> Usually, Rayleigh fading means that the envelope of the channel
>> impulse response follows a Rayleigh function.  A Rayleigh function
>> is zero at t=0, reaches a peak at some non-zero t, then declines
>> to zero at t=infinity.  Within this envelope, the number,
>> density, and pattern of individual impulses are further parameters
>> of your channel model.  A Rayleigh channel might naturally arise
>> from certain ensembles of large numbers of reflections, but
>> this is not the only possible type of Rayleigh channel.
>>
> 
> 
> This is a very interesting meaning for Rayleigh fading
> that I have not encountered before. A reference where
> more details can be found would be much appreciated.
> 


I think he is mixing up the impulse response with the probability 
density function.

Greetz,
Sebastian Doht

cpshah99 <cpshah99@rediffmail.com> wrote:

>Hello All
>
>Thanks a lot for your replies.
>
>But I am still confused from the simulation point of view.

>After your replies, I am more confused. 

Please ignore my previous reply, as I apparently am not using the
terms correctly.

Steve

>Please ignore my previous reply, as I apparently am not using the
>terms correctly.
>
>Steve
>

Hi Steve

Is that it? Please help me understand this. I am going mad over this
topic.

Chintan

On Jan 24, 11:10&#4294967295;am, "cpshah99" <cpsha...@rediffmail.com> wrote:
> >Please ignore my previous reply, as I apparently am not using the
> >terms correctly.
>
> >Steve
>
> Hi Steve
>
> Is that it? Please help me understand this. I am going mad over this
> topic.
>
> Chintan

Chintan, with all due respect I think you should spend some time with
the basics.  I respect very much your eagerness to learn, but you seem
to be discovering new holes on the hull of your boat whenever you
think
that you can start pedaling.

How about an actually systematic learning effort?  Here's a book I can
recommend:

http://www.eecs.berkeley.edu/~dtse/book.html

Please read their first few chapters carefully.

In particular, you seem to be confusing not only terminologies, but
more importantly how to apply physical and mathematical models into
simulation, or dryly speaking computer representation.

This is NOT easy at all.  In fact, you can see from many posts here
that there are many "drive-by postings" on this topic.  And further,
you are almost the only one who's persisting, which is to your
credit :-).

So don't give up man, but please change your strategy and learn the
topic systematically instead of a la carte :-).

This is a fun topic though, and quite useful to cases other than
wireless
communication.  It's a good way to learn how to do computer modeling
of a (stochastic) physical phenomena.

Cheers,
Julius

Sebastian Doht  <seb_doht@lycos.com> wrote:

>dvsarwate@yahoo.com schrieb:

>> This is a very interesting meaning for Rayleigh fading
>> that I have not encountered before. A reference where
>> more details can be found would be much appreciated.

>I think he is mixing up the impulse response with the probability 
>density function.

Yes, you're right.  I had that confused.  The envelope of the 
impulse response can be whatever shape is desired by the modeler.  
What is common is for the magnitude of the individual impulses to 
have a Rayleigh distribution.

Looking at some notes, I see this was used by Turin in 1972
("A statistical model of urban multipath propogation", Trans.
Vehicular Technology), and more recently in the
Saleh-Valenzuela and Spencer models which are commonly used
(at least in some applications).

All of the above are static channel models.

Steve