[USRP-users] USRP captured data information

Marcus Müller marcus.mueller at ettus.com
Wed Apr 15 07:57:51 EDT 2015

I'm not familiar with optical models myself, so this is only a guess:
*Geometrical* optical model sounds like it would ignore all effects that
happen due to interference, meaning that you wouldn't see any multipath
fading; I'm pretty sure that would not be the model of choice for radio
waves inside a subway, since they definitely will de- and constructively

When I was proofreading a few pages of a thesis, I skimmed through [1]
to get a understanding for the topic; maybe that will give you some
inspiration for what to measure; it sadly is already more than 20 years
old, so there might be "cooler" existing models that I don't know of.
Anyway, it really nicely covers the three main aspects that you probably
also will need to cover:

1. it describes how the channel is modelled (as a discrete impulse response)
2. it describes why considering the channel as something that varies
randomly is necessary and thus explains the need for statistical description
3. it is kind of a lengthy tutorial on how to actually characterize the
channel. I didn't read that part completely, because it was "a bit too
long", but then again, I was not in the business of modelling a channel

I must admit that I think what you're doing might be a bit hard for
someone completely new to radio communications and if that actually is
the case for you, you might find it hard to understand all the technical
terms in the first three chapters of [1], but you only said you were new
to SDR, so my hope is that you already know most of this stuff, or will
have fun catching up on a bit of channel theory.


[1] http://wsl.stanford.edu/~ee359/hashemi_indoor.pdf
On 04/15/2015 01:33 PM, alok ranjan wrote:
> Dear Marcus,
> Thank you for your all valuable suggestions and assistance to my issues.
> Yes! you are very true that the subway scenario suffers from multi
> path effects and shadowing phenomena.
> I have just completed my Graduation and working as project assistant.
> As such i dont have exact idea that which channel model will work
> fine. I am reading literatures on wireless communication
> possibilities  and found that Geometrical Optical model may work fine
> and there are literatures available.
> What you have  suggested, I will keep in mind for sure about the
> antenna and the filters.
> If you have any other suggestions then I will feel happy to hear from you.
> Thank you, Regards,
> Alok Ranjan
> ------------------------------------------------------------------------
> Date: Wed, 15 Apr 2015 11:49:17 +0200
> To: usrp-users at lists.ettus.com
> Subject: Re: [USRP-users] USRP captured data information
> From: usrp-users at lists.ettus.com
> Hi Alok,
> we'll, Matlab is a math tool, and thus, you'll have to find a
> real-valued mathematical representation of the entities you want to plot.
> Power of a digital signal is commonly defined as the square magnitude
> of the samples, that is $\Re\{z\}^2+\Im\{z\}^2 = {|z|}^2$, so this
> might be what you want to plot.
> However, you might find yourself in a situation where you'll first
> want to define more closely what you're measuring. Do you want to
> measure the RX power of a single sine wave? In that case, it would be
> advisable to use a very narrowband digital filter, just to get as
> little power as possible into your measurement. If you transmit
> digital data (most probably what you do when using benchmark_tx), you
> might now your TX pulse shape, and thus you can use a matched filter
> on the RX side, giving you the maximum SNR any filter could offer.
> Also, the frequency range from 300MHz to 5GHz is quite large -- unless
> you have an extremely broadband antenna, you will need to use
> different antennas. And: every antenna has a frequency-dependent
> behaviour as well, making both the geometrical properties and the
> maximum achievable field strength extraction a function of the carrier
> frequency. You will be fine if you use "simple" geometries like
> monopole antennas and for small frequency deviations, and if you try
> to keep RX and TX in each others main direction, but that's another
> factor to consider; so I'd recommend doing one center frequency first,
> shifting your signal in smaller steps around that, and then trying out
> other frequencies.
> Your application really sounds exciting; subway channels really sound
> like they'd actually have impressive multipath propagation, small
> scale fading etc, due to being in a metal cage; this together with the
> doppler of signals leaving the cabin, bouncing of the tunnel seems
> really interesting! Do you have a reference channel model you're
> building on? Maybe the Matlab IEEE802.11n indoor channels?
> Greetings,
> Marcus
> On 04/15/2015 07:17 AM, alok ranjan via USRP-users wrote:
>     Dear Marcus and Raj,
>     Thank you for your assistance so far.
>     I am trying to do some channel modeling for wireless communication
>     in subway.  To accomplish this I am using two USRP1 having DB WBX
>     connected to my pc over USB 2.0.
>     As I am new to both SDR and USRP , I initially used the two
>     scripts i.e. benchmark_tx.py and benchmark_rx.py for transmission
>     and reception.  I am sending baseband samples at different
>     frequencies and modulation scheme at varying amplitude of TX. (The
>     distance between the TX and RX was 5 mtr)
>     I have captured the transmitted data at different frequencies at
>     receiver using uhd_rx_cfile.  As, the Marcus has mentioned that
>     these captured data is in terms of I/Q complex format, 16 bit for
>     each. 
>     I am trying to simply measure the power in the signal i have
>     recorded at receiver using uhd_rx_cfile. I want to do this
>     analysis at different frequencies say 300 MHz to 5 GHz. 
>     Another things i would like to do is capture the data with the
>     moving receiver while the transmitter is fixed.
>     Now, I want to do some curve fitting (best fit, polynomial or
>     linear regression) on the captured data therefore I converted the
>     captured data into .mat format using octave. When i load the data
>     in workspace in MATLAB, I am not getting how to start with the
>     plot and do curve fitting for Recorded reception power level VS
>     distance to further proceed for channel modeling
>     Any help in this regard is appreciated.
>     Thank you all for your time and endeavour  pain.
>     Best Regards,
>     Alok Ranjan
>     Graduate student
>     NIT, Rourkela
>     ------------------------------------------------------------------------
>     Date: Tue, 14 Apr 2015 15:32:57 -0400
>     To: usrp-users at lists.ettus.com <mailto:usrp-users at lists.ettus.com>
>     Subject: Re: [USRP-users] USRP captured data information
>     From: usrp-users at lists.ettus.com <mailto:usrp-users at lists.ettus.com>
>     Alok,
>     If you tell us a bit more about what you are trying to accomplish
>     with your curve fitting and plotting, maybe we can be of more
>     assistance. What kind of curve fit or plot do you wish to
>     generate? Are you trying to demodulate a signal? Analog or
>     digital? Are you trying to simply measure the power in the signal
>     you recorded? What center frequency and bandwidth are you
>     interested in recording? What kinds of signals are being
>     transmitted in that band? Are you going to move your receiver
>     around and take measurements from fixed transmitters?
>     -- 
>     Raj Bhattacharjea
>     Georgia Institute of Technology
>     School of Electrical and Computer Engineering
>     http://www.prism.gatech.edu/~gtg037s/
>     <http://www.prism.gatech.edu/%7Egtg037s/>
>     404.894.7516
>     _______________________________________________ USRP-users mailing
>     list USRP-users at lists.ettus.com
>     <mailto:USRP-users at lists.ettus.com>
>     http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>     _______________________________________________
>     USRP-users mailing list
>     USRP-users at lists.ettus.com <mailto:USRP-users at lists.ettus.com>
>     http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> _______________________________________________ USRP-users mailing
> list USRP-users at lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150415/e2808328/attachment-0002.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 1467 bytes
Desc: not available
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150415/e2808328/attachment.png>

More information about the USRP-users mailing list