DSRC demodulation with gr-ieee802_11

GNU Radio
1

I’m new to Gnu Radio but I’m asked to demodulate a DSRC signal and I
don’t
know how handle that. DSRC is used in Intelligent Vehicule Systems, it
is
modulated with OFDM int the 5.9Hz band and is based on the IEEE 802.11p
standard.

On my first approach, I found some papers about the module
gr-ieee802_11:
https://archive.fosdem.org/2014/schedule/event/towards_an_open_source_ieee_802_11p_stack/

I installed the module as described here:
GitHub - bastibl/gr-ieee802-11: IEEE 802.11 a/g/p Transceiver and tried to use the
examples/wifi_rx.grc file to start. Here begin my problem: using the
same
record_source_file, I have several results printed on my console. It’s
seems that the module cannot detect OFDM frames.

Is there something I missed to use gr-ieee802_11 module ?

I’m using Gnu Radio 3.7.2 with USRP N210, my record_source_file is
sampled
with 5Msps at 5.127 GHz.

Do not hesitate to ask if I missed some required information.

Thanks !

2

Hi,

On 09/22/2014 03:51 PM, Jean-Baptiste Truffault wrote:

I installed the module as described here:
GitHub - bastibl/gr-ieee802-11: IEEE 802.11 a/g/p Transceiver and tried to use the
examples/wifi_rx.grc file to start. Here begin my problem: using the
same record_source_file, I have several results printed on my console.
It’s seems that the module cannot detect OFDM frames.

I’m afraid you have to be a bit more verbose… What did you do? And
what exactly went wrong?

Did you follow the steps in the readme?

I’m using Gnu Radio 3.7.2 with USRP N210, my record_source_file is
sampled with 5Msps at 5.127 GHz.

IIRC, you need at least 3.7.3 (will add that to the readme).

Best,
Bastian

3

Sorry for the time, I had some trouble with the 3.7.3 install, finally I
installed the 3.7.5 thanks to the script
http://www.sbrac.org/files/build-gnuradio.

First, I recorded a DSRC signal thanks to Gnu Radio and my USRP N210
(with
a CBX daughter card): 5Msps, 5.127 GHz. I now use this file as source.
It
contends some DSRC exchanges that I would like to decode (I am quite
sure
that the record is good thanks to the shape of it, cf attached picture
[3_fft.png]).

My grc scheme [1_grc1.png & 2_grc2.png] is very similar of yours, I just
attached it to be sure.

The problem is when I read the debug from the OFDM Decode Mac,
at the first time I launch the script, I get something like:

80 be 44 b5 e1 71 e5 57 31 ac 58 4e 0b db b8 1f
03 63 c7 b4 1b e5 2c 8c 76 da 20 f5 c2 e0 8e bc
62 ad 12 37 d5 07 1a 5f e6 d4 03 a4 e7 db 1c 9c
eb 99 1c 52 e8 83 13 06 ac 08 70 4b bf 71 0c ab
74 52 2c fc f9 ca 05 4a 29 55 c9 c6 cb 7c f8 51
22 60 6f 5f 64 d7 55 f9 32 f6 79 7c 61 ea 89 86
f3 33 ea 85 43 52 12 6e b2 92 e7 3c 81 5c 49 f2
b6 30 03 dc c4 15 f4 cd f0 a1 5a 1b c9 88 a8 94 …

the second time I launch the script, I get something completly different
that looks like:

00 ee 2b f8 e9 7c d4 5e 93 22 e9 fc d4 1a 9e 1a
d5 3e 03 0b 93 da 15 fe 3a 2f 53 fd c0 9c 98 a6
9e 8f 8c 0f 30 cc f6 a6 16 46 81 54 b7 95 3a 1e
20 36 20 b6 91 3c 20 55 ab 02 ea a4 9f e8 69 88
97 77 c8 de 92 25 0e 18 6b e1 29 4f e6 3c 86 65
b0 7a 61 0b 47 d6 22 38 19 4c 0b b9 fb 1b 53 83
aa 4b d7 fe 9f 4d de 0f 8b 99 a7 de a4 09 0d d5
1c d5 6c dc 66 bf 5f 7c dc b7 a3 fa 88 b3 c8 7b
11 1b 04 41 1a 41 5d 51 74 ea 48 11 33 d2 d1 82 …

Those results should be the same because they come from the same record
file…

The debug indicates that the checksum is wrong so the frame is dropped
and
I do not receive anything on the OFDM Parse MAC Block.

My first idea is that it comes from the OFDM frame detection because of
the
strange shape of it (cf [5_autocorrelation.png]), it acts like there
were
always something to decode when the messages come with a burst on the
FFT
[3_fft.png].

Another thing that is different from your result is the constellation
scope
that may have 4 clusters (instead of one on the 4_constellation.png)

I would be very kind if you had some idea where I did something wrong !

Thanks in advance for your time and your help !

2014-09-23 4:16 GMT+02:00 Bastian B. [email protected]:

4

On 09/24/2014 04:14 PM, Jean-Baptiste Truffault wrote:

Sorry for the time, I had some trouble with the 3.7.3 install, finally I
installed the 3.7.5 thanks to the script
http://www.sbrac.org/files/build-gnuradio.

First, I recorded a DSRC signal thanks to Gnu Radio and my USRP N210
(with a CBX daughter card): 5Msps, 5.127 GHz. I now use this file as
source. It contends some DSRC exchanges that I would like to decode (I
am quite sure that the record is good thanks to the shape of it, cf
attached picture [3_fft.png]).

I don’t think so. This does not look like a 10MHz OFDM signal…

My grc scheme [1_grc1.png & 2_grc2.png] is very similar of yours, I just
attached it to be sure.

This looks OK.

My first idea is that it comes from the OFDM frame detection because of
the strange shape of it (cf [5_autocorrelation.png]), it acts like there
were always something to decode when the messages come with a burst on
the FFT [3_fft.png].

The autocorrelation is always high so frame detection is triggered all
the time. This might happen when you start the USRP source. Try to drop
samples at startup (skip head block).

Another thing that is different from your result is the constellation
scope that may have 4 clusters (instead of one on the 4_constellation.png)

Since you are not synced on a real frame you just see noise.

Best,
Bastian

5

OK, I try to record a new signal and I tell you if I get something new.

2014-09-24 17:51 GMT+02:00 Bastian B. [email protected]:

6

I see there some double used terms. Is it the European ETSI style DSRC,
or the US WAVE style? 802.11p?

Ralph.

From: discuss-gnuradio-bounces+ralph=removed_email_address@domain.invalid
[mailto:discuss-gnuradio-bounces+ralph=removed_email_address@domain.invalid] On Behalf Of
Jean-Baptiste Truffault
Sent: Thursday, September 25, 2014 9:54 AM
To: Bastian B.
Cc: [email protected]
Subject: Re: [Discuss-gnuradio] DSRC demodulation with gr-ieee802_11

OK, I try to record a new signal and I tell you if I get something new.

2014-09-24 17:51 GMT+02:00 Bastian B. <[email protected]
mailto:[email protected] >:

On 09/24/2014 04:14 PM, Jean-Baptiste Truffault wrote:

Sorry for the time, I had some trouble with the 3.7.3 install, finally I
installed the 3.7.5 thanks to the script
http://www.sbrac.org/files/build-gnuradio.

First, I recorded a DSRC signal thanks to Gnu Radio and my USRP N210
(with a CBX daughter card): 5Msps, 5.127 GHz. I now use this file as
source. It contends some DSRC exchanges that I would like to decode (I
am quite sure that the record is good thanks to the shape of it, cf
attached picture [3_fft.png]).

I don’t think so. This does not look like a 10MHz OFDM signal…

My grc scheme [1_grc1.png & 2_grc2.png] is very similar of yours, I just
attached it to be sure.

This looks OK.

My first idea is that it comes from the OFDM frame detection because of
the strange shape of it (cf [5_autocorrelation.png]), it acts like there
were always something to decode when the messages come with a burst on
the FFT [3_fft.png].

The autocorrelation is always high so frame detection is triggered all
the time. This might happen when you start the USRP source. Try to drop
samples at startup (skip head block).

Another thing that is different from your result is the constellation
scope that may have 4 clusters (instead of one on the
4_constellation.png)

Since you are not synced on a real frame you just see noise.

Best,
Bastian

7

Finally I’ve got something better concerning the autocorrelation, but
it’s
still very noisy on the constellation…
maybe something I’m still doing wrong ?

2014-09-25 13:05 GMT+02:00 Bastian B. [email protected]:

8

Hi Ralph,

On 25 Sep 2014, at 11:10, Ralph A. Schmid, dk5ras [email protected]
wrote:

I see there some double used terms. Is it the European ETSI style DSRC, or the
US WAVE style? 802.11p?

IEEE WAVE and ETSI ITS G5 are both based on IEEE 802.11p. The module
implements the phy and parts of the mac, which are defined in 11p. So it
doesnt matter what you use it for.

Best,
Bastian

9

On 25 Sep 2014, at 15:07, Jean-Baptiste Truffault
[email protected] wrote:

Finally I’ve got something better concerning the autocorrelation, but it’s still
very noisy on the constellation…
maybe something I’m still doing wrong ?

Now it looks like you are just concatenating long and short preamble and
miss the actual frame.

Best,
Bastian

10

is there a way to troubleshoot that kind of problem ?

2014-09-25 15:47 GMT+02:00 Bastian B. [email protected]:

11

On 25 Sep 2014, at 15:58, Jean-Baptiste Truffault
[email protected] wrote:

is there a way to troubleshoot that kind of problem ?

I think you should read the readme and follow the steps in order to find
out what works and at which step things break (for example I dont know
if the loopback flow graph works already).

If it does not work over the air use graphical sinks and the log / debug
options of the blocks to find out at which stage of the receiver things
go wron

In case you still have problems you should describe your setup in
greater detail (including how you generate your frames).

Hope it helps,
Bastian

12

On Thu, Sep 25, 2014 at 09:53:54AM +0200, Jean-Baptiste Truffault wrote:

First, I recorded a DSRC signal thanks to Gnu Radio and my USRP N210
(with a CBX daughter card): 5Msps, 5.127 GHz.

I don’t think so. This does not look like a 10MHz OFDM signal…

Are you trying to capture a 10 MHz wide signal at 5 Msps? That would be
a problem.

Mike