Choosing start of frame bits, studies?

GNU Radio
1

Hi all,

The gr_correlate_access_code_bb block takes a 64-bit access code as
input to detect the start of incoming frames using a hamming distance
with a threshold. The packet utilities has a default access code and
uses a threshold of 12.

I’m not sure who wrote the code and who picked the default access code
and threshold, but my instinct tells me Eric :slight_smile:

Was the access code picked randomly, or was there some basic rule of
thumb for generating it? What about the threshold of 12? What about
choosing a 64bit access as opposed to 32bit?

Does anyone know of any studies done on these variables that I could
read about? They have obvious impacts on false positives and negatives.

When it comes down to fully decoding the frame, false positives can
easily be picked out by examining the frame header. But, I’m working on
something that is attempting to perform actions without fully decoding
the frame header and frame, so the access code itself has a strong
impact.

Thanks!
George

2

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

George N. wrote:

Was the access code picked randomly, or was there some basic rule of
thumb for generating it? What about the threshold of 12? What about
choosing a 64bit access as opposed to 32bit?

Standards for access codes are interesting. The Barker spreading code
use in 802.11 1 and 2Mbps rates, for instance, is one of a set of
optimal codes where B (dot) B = |B| while
B (dot) (B rotated by any nonzero number of bits) = 1. This means that
if you’re sliding a correlator along a set of Barker-modulated bits, you
get strong peaks at the correct offset and (basically) nothing at any
other offset.

Usually access codes have a property somewhat like that. I think the
nuances of this choice are strongly tied in, however, with what
modulation scheme you’re using.

  • -Dan
    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)
    Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHh5/0y9GYuuMoUJ4RAny5AKCNQy0t5SY+sD0qwg5qpLfK9/jhmgCeLyDC
0Wf5PDIkwM8GDUZY0Aa0cmM=
=z/qB
-----END PGP SIGNATURE-----

3

Dan H. wrote:

Standards for access codes are interesting. The Barker spreading code
use in 802.11 1 and 2Mbps rates, for instance, is one of a set of
optimal codes where B (dot) B = |B| while
B (dot) (B rotated by any nonzero number of bits) = 1. This means that
if you’re sliding a correlator along a set of Barker-modulated bits, you
get strong peaks at the correct offset and (basically) nothing at any
other offset.

In addition to the auto-correlation properties described above, there
may be a requirement for a set of multiple access codes to have
negligible cross-correlation with other members of the set. Some
protocols use multiple access codes (also known as sync codes) to signal
the type of frame that follows (e.g., the FLEX pager protocol, see
pageri_flex_codes.cc in the tree).

–
Johnathan C.
Corgan Enterprises LLC
http://corganenterprises.com