Forums

input data to check MUSIC algorithm

Started by krish_dsp September 1, 2008
Hi everyone,
I am implementing MUSIC algorithm for DOA estimation through MATLAB and
DSP. For checking my algorithm I need a simulated input that resembles a
multipath fading channels. Friend of mine told me that there are models in
MATLAB which simulate these data. 
I searched in MATLAB and i got a simulink model of multi path fading
channel by rayleigh and rician fading. But I am not able to get the output
of the model as a text or ram file. All i see is the animation windows. I
am not well exposed to simulink.
Can anyone help me in this? If there are other ways to simulate the
wirless channel, please throw light on that.
Thank you.
On 2 Sep, 02:29, "krish_dsp" <nmkr...@gmail.com> wrote:
> Hi everyone, > I am implementing MUSIC algorithm for DOA estimation through MATLAB and > DSP. For checking my algorithm I need a simulated input that resembles a > multipath fading channels.
First things first: Did you verify the algorithm with perfect data? That is, a single source under perfect conditions? Did you add noise to see how the algorithm works under different SNRs? Added more sources? There is lots of work to be done before you can get to the specifics of particular scenarios. Rune
Thanks rune...
I have not completed my music algorithm and I started worryin abt the
input signal.
Now I have another doubt. What exactly does a steering vector means. For
an array of N sensors, the steering vector is given by a N x1 matrix of
[1,exp(-j*2*pi*f*del*1),exp(-j*2*pi*f*del*2),......exp(-j*2*pi*f*del*N-1)]
where del is the time delay between the wavs incident on consecutive
sensors given by del = d * sin(theta)/c, where d is the sensor seperation
and c = speed of light (not sure abt that!!!)
This is my underdtanding but i am not able to correlate the concept of
'steering vector' with the above formulae...Can u guide me on that?

Also, can i use the above formulae letter by letter in matlab? Or should I
make any changes. I am havinf this question b'cos seperation between arrays
can be in the order of meters but c is about 10^8 meters. I have a feeling
that if I use the formula as it is above, it wont be correct in matlab.


On 3 Sep, 07:49, "krish_dsp" <nmkr...@gmail.com> wrote:
> Thanks rune... > I have not completed my music algorithm and I started worryin abt the > input signal. > Now I have another doubt. What exactly does a steering vector means. For > an array of N sensors, the steering vector is given by a N x1 matrix of > [1,exp(-j*2*pi*f*del*1),exp(-j*2*pi*f*del*2),......exp(-j*2*pi*f*del*N-1)] > where del is the time delay between the wavs incident on consecutive > sensors given by del = d * sin(theta)/c, where d is the sensor seperation > and c = speed of light (not sure abt that!!!) > This is my underdtanding but i am not able to correlate the concept of > 'steering vector' with the above formulae...Can u guide me on that?
Sorry no, I can't help. The only thing which will work is that you sit down and work your way through the details, one by one.
> Also, can i use the above formulae letter by letter in matlab? Or should I > make any changes. I am havinf this question b'cos seperation between arrays > can be in the order of meters but c is about 10^8 meters. I have a feeling > that if I use the formula as it is above, it wont be correct in matlab.
Once you understand what is going on it will not matter what is 'correct in matlab' since you will know what is correct 'on paper' and will be able to translate that into matlab code. Rune