DOA using MUSIC and Circular antenna array

HI Guys

I do not have much experience on this topic, I would appreciate some
guidance.

I have read about MUSIC and DOA using linear array. It is easy to
understand the topic.

Is there any good paper/book which provides treatment of DOA and MUSIC
using circular array.

Many thanks.

Best Regards

Chintan

On Nov 3, 9:22&#4294967295;am, "cpshah99" <cpshah99@n_o_s_p_a_m.rediffmail.com>
wrote:
> HI Guys
>
> I do not have much experience on this topic, I would appreciate some
> guidance.
>
> I have read about MUSIC and DOA using linear array. It is easy to
> understand the topic.
>
> Is there any good paper/book which provides treatment of DOA and MUSIC
> using circular array.

It depends on what version of MUSIC you read up on. If you used
the classical version from Schmidt's late '70s papers, the extension
to circular arrays is straight-forward. If you used any other
version,
you will need to read up on classical MUSIC.

Rune

>On Nov 3, 9:22=A0am, "cpshah99" <cpshah99@n_o_s_p_a_m.rediffmail.com>
>wrote:
>> HI Guys
>>
>> I do not have much experience on this topic, I would appreciate some
>> guidance.
>>
>> I have read about MUSIC and DOA using linear array. It is easy to
>> understand the topic.
>>
>> Is there any good paper/book which provides treatment of DOA and MUSIC
>> using circular array.
>
>It depends on what version of MUSIC you read up on. If you used
>the classical version from Schmidt's late '70s papers, the extension
>to circular arrays is straight-forward. If you used any other
>version,
>you will need to read up on classical MUSIC.
>
>Rune
>

Thanks Rune.

I read MUSIC from a book ''Narrowband Direction of Arrival Estimation for
Antenna Arrays (Synthesis Lectures on Antennas)''.

If you can give me the title of the paper, that will be great.

Best Regards

Chintan

On Nov 3, 10:07&#4294967295;am, "cpshah99" <cpshah99@n_o_s_p_a_m.rediffmail.com>
wrote:
> >On Nov 3, 9:22=A0am, "cpshah99" <cpshah99@n_o_s_p_a_m.rediffmail.com>
> >wrote:
> >> HI Guys
>
> >> I do not have much experience on this topic, I would appreciate some
> >> guidance.
>
> >> I have read about MUSIC and DOA using linear array. It is easy to
> >> understand the topic.
>
> >> Is there any good paper/book which provides treatment of DOA and MUSIC
> >> using circular array.
>
> >It depends on what version of MUSIC you read up on. If you used
> >the classical version from Schmidt's late '70s papers, the extension
> >to circular arrays is straight-forward. If you used any other
> >version,
> >you will need to read up on classical MUSIC.
>
> >Rune
>
> Thanks Rune.
>
> I read MUSIC from a book ''Narrowband Direction of Arrival Estimation for
> Antenna Arrays (Synthesis Lectures on Antennas)''.
>
> If you can give me the title of the paper, that will be great.
>
> Best Regards
>
> Chintan- Hide quoted text -
>
> - Show quoted text -

Reference 2 on this page:

http://www.mathworks.com/help/toolbox/signal/pmusic.html

You might also find it useful to read the intro to Roy & Kailath's
paper on ESPRIT from around 1989. They point out a few issues with
MUSIC an DoA estimation you might want to be aware of.

Rune

>
>Reference 2 on this page:
>
>http://www.mathworks.com/help/toolbox/signal/pmusic.html
>
>You might also find it useful to read the intro to Roy & Kailath's
>paper on ESPRIT from around 1989. They point out a few issues with
>MUSIC an DoA estimation you might want to be aware of.
>
>Rune
>

Thanks Rune.

I will go thru this papers.

Regards

Chintan

Hi Rune

As you had said, the theory of UCA is straight forward.

I have done some simulations for one signal.

So I simulated ULA and MUSIC. This seems to work fine. Basically, we have
to search for power in each azimuth angle's direction and select the one
with maximum.

And same for UCA and MUSIC, but now for a given elevation angle, we have to
search for all azimuth angles. Is this correct?

However, the problem that I have is UCA-MUSIC fails when I add noise even
at very high SNR but I get exact results for noise-less case.

ULA-MUSIC works fine even when I add noise. Of course the performance
degrades for low SNR.

What could be wrong?

Chintan

On Nov 9, 11:17&#4294967295;am, "cpshah99" <cpshah99@n_o_s_p_a_m.rediffmail.com>
wrote:
> Hi Rune
>
> As you had said, the theory of UCA is straight forward.
>
> I have done some simulations for one signal.
>
> So I simulated ULA and MUSIC. This seems to work fine. Basically, we have
> to search for power in each azimuth angle's direction and select the one
> with maximum.
>
> And same for UCA and MUSIC, but now for a given elevation angle, we have to
> search for all azimuth angles. Is this correct?

I haven't done that so I wouldn't know, but yes - that's an
obvious first attempt.

> However, the problem that I have is UCA-MUSIC fails when I add noise even
> at very high SNR but I get exact results for noise-less case.
>
> ULA-MUSIC works fine even when I add noise. Of course the performance
> degrades for low SNR.
>
> What could be wrong?

I have no idea. These things could be highly nonlinear or
sensitive to noise - or both.

Does the azimuth-only version work? If so, look for sensitivities
in elevation. And remember, these kinds of things are ill-posed
at the outset: What azimuth reading would you expect if the
source is located vertically above the center of the circular
array?

Rune

>Does the azimuth-only version work? If so, look for sensitivities
>in elevation. 

It does not work. UCA-MUSIC completely fails even for very low noise
variance. 

>And remember, these kinds of things are ill-posed
>at the outset: What azimuth reading would you expect if the
>source is located vertically above the center of the circular
>array?

In this case I would expect the azimuth to be 0.

Chintan