>
>mrlius@yahoo.it wrote:
>> Hello all,
>> I have to realize a Smart Antenna system using the LMS algorithm on an
>> uniform linear array.
>> Let's suppose that we know already the angles of arrival of the
desired
>> signal and of one or more interferers.
>> How can I calculate the LMS weights if I don't know how exactly is the
>> desired signal?

   In this state u should decide that u want to use :
1) A blind method, or 
2) A training-based method

>> I can compute the weights knowing the desired signal but then it would
>> be useless to calculate the weights and to use this method since the
>> desired signal is what I don't know and is what I'm looking to receive
>> and filter for (I know only the angle of arrival).


If u use the 2nd method, u should consider training sequences between ur
transmited signal ( that acts as ur DESIRED in the LMS algorithm) and
after adapting the weights u will detect the unknown.

If u want to set the weights blindly(without transmitting any data), tell
me to explain more.

regards
Morteza

mrlius@yahoo.it wrote:
> Hello all,
> I have to realize a Smart Antenna system using the LMS algorithm on an
> uniform linear array.
> Let's suppose that we know already the angles of arrival of the desired
> signal and of one or more interferers.
> How can I calculate the LMS weights if I don't know how exactly is the
> desired signal?
> I can compute the weights knowing the desired signal but then it would
> be useless to calculate the weights and to use this method since the
> desired signal is what I don't know and is what I'm looking to receive
> and filter for (I know only the angle of arrival).
> Thanks.


I suggest you look at R. T. Compton, Jr.'s book "Adaptive Antennas -
Conceots and Performance".  The section on page 6 is titled "The LMS
Adaptive Array".  In my opinion, this is one of the best books on
adaptive antennas.


Maurice Givens

>
>mrlius@yahoo.it skrev:
>> Hello all,
>> I have to realize a Smart Antenna system using the LMS algorithm on an
>> uniform linear array.
>> Let's suppose that we know already the angles of arrival of the
desired
>> signal and of one or more interferers.
>> How can I calculate the LMS weights if I don't know how exactly is the
>> desired signal?
>
>You need to find a set of weights that maximizes the sensitivity in the
>
>direction of the desired signal, and suppresses everything else. This
>is
>a standard problem in adaptive beamforming, although I don't remember
>the exact term for it.
>
>Rune
>
>

I think its called "null-steering".  You place nulls in the direction of
unwanted signals, and a beam in the direction of the desired signal.

<mrlius@yahoo.it> wrote in message 
news:1162606389.804126.168890@m7g2000cwm.googlegroups.com...
> Hello all,
> I have to realize a Smart Antenna system using the LMS algorithm on an
> uniform linear array.
> Let's suppose that we know already the angles of arrival of the desired
> signal and of one or more interferers.
> How can I calculate the LMS weights if I don't know how exactly is the
> desired signal?
> I can compute the weights knowing the desired signal but then it would
> be useless to calculate the weights and to use this method since the
> desired signal is what I don't know and is what I'm looking to receive
> and filter for (I know only the angle of arrival).
> Thanks.
>

You didn't say if you know the frequency or frequency range of the "desired" 
signal.

You did say you know the angle of arrival of the desired signal.

If the desired signal is at a single frequency or if the desired signal is 
at least such that it's more or less in some frequency band then the 
beamforming weights should be determinable without adaptation.

If you don't know the desired signal but *do* know the undesired signal then 
you might implement something like an adaptive line canceller.

Maybe something like this would work:?

1) Steer fixed beams of the array in the direction of the desired signals 
and the interferers.

2) Use the output of the beams pointed at the interferers as a reference 
source in an adaptive line canceller - to remove remnants of the interferers 
in the beams pointed toward the desired signals.  To the extent that the 
"interference" beams pick up the desired signals, this will also attenuate 
the desired signals to a degree.  So, it will be necessary that there be 
some signal-to-interference or interference-to-signal gain in both beam 
sets.

Fred 

mrlius@yahoo.it wrote:

> Hello all,
> I have to realize a Smart Antenna system using the LMS algorithm on an
> uniform linear array.
> Let's suppose that we know already the angles of arrival of the desired
> signal and of one or more interferers.
> How can I calculate the LMS weights if I don't know how exactly is the
> desired signal?
> I can compute the weights knowing the desired signal but then it would
> be useless to calculate the weights and to use this method since the
> desired signal is what I don't know and is what I'm looking to receive
> and filter for (I know only the angle of arrival).
> Thanks.

You need to design the antenna in the way to maximize the ratio of the 
desired signal to the interferrence. This is a typical problem of 
numeric optimization. You can use LMS or any other multidimensional 
optimization algorithm.

Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

mrlius@yahoo.it skrev:
> Hello all,
> I have to realize a Smart Antenna system using the LMS algorithm on an
> uniform linear array.
> Let's suppose that we know already the angles of arrival of the desired
> signal and of one or more interferers.
> How can I calculate the LMS weights if I don't know how exactly is the
> desired signal?

You need to find a set of weights that maximizes the sensitivity in the

direction of the desired signal, and suppresses everything else. This
is
a standard problem in adaptive beamforming, although I don't remember
the exact term for it.

Rune

Hello all,
I have to realize a Smart Antenna system using the LMS algorithm on an
uniform linear array.
Let's suppose that we know already the angles of arrival of the desired
signal and of one or more interferers.
How can I calculate the LMS weights if I don't know how exactly is the
desired signal?
I can compute the weights knowing the desired signal but then it would
be useless to calculate the weights and to use this method since the
desired signal is what I don't know and is what I'm looking to receive
and filter for (I know only the angle of arrival).
Thanks.