OpenBCM V1.13 (Linux)

Packet Radio Mailbox

DB0FHN

[JN59NK Nuernberg]

 Login: GUEST





  
ZL3AI  > APRS     07.11.00 07:39l 217 Lines 8026 Bytes #999 (0) @ WW
BID : 34344_ZL3AI
Read: GUEST
Subj: TAPR Digest, Nov 04, 2/4
Path: DB0MRW<DB0RGB<OE5XBL<OE2XUM<OE5XBR<OE3XBS<OE6XYG<9A0BBS<9A0YRB<VK6BBS<
      VK2CPG<VK2CZR<VK2IO<VK2GJB<ZL2TZE<ZL2TZE<ZL2WA<ZL2BAU<ZL3VML
Sent: 001105/1714Z @:ZL3VML.#80.NZL.OC #:46305 [Chch-NZ] FBB7.00g25 $:34344_ZL3
From: ZL3AI@ZL3VML.#80.NZL.OC
To  : APRS@WW


after looking at the syntax of the URL for 10 or 15 seconds, it was
intuitive to me...

73, gerry
--
Gerry Creager -- gerry@cs.tamu.edu
Network Engineering			|Research focusing on
Academy for Advanced Telecommunications |Satellite Geodesy and 
and Learning Technologies		|Geodetic Control
Texas A&M University	979.458.4020  (Phone) -- 979.847.8578  (Fax)

----------------------------------------------------------------------

Subject: RE: FINDU Status/comment
From: "Brian B. Riley" <brianbr@macconnect.com>
Date: Sat, 4 Nov 2000 11:01:54 -0500
X-Message-Number: 9



> -----Original Message-----
> From: Gerry Creager N5JXS
> Sent: Saturday, November 04, 2000 10:20 AM

> Steve Sampson wrote:
> > Completely undocumented.
> > ----- Original Message -----
> > > Already there.  Try:
> http://www.findu.com/cgi-bin/rawposit.cgi?call=k9sa-9&start=24
> &length=12
> > >
> > > Obviously insert your call and the start time and length.
> > > Example starts 24 hours ago and runs for 12 hours.
>
> Er, no... Steve documented it on the list some time back.
> Like a lot of code-swingers I know, he does more coding than
> documenting.  However, after looking at the syntax of the
> URL for 10 or 15 seconds, it was intuitive to me...

  If you want to see the bulk of the calls that are available, go to my
APRS Query page at

  <http://www.wulfden.org/APRSQuery.shtml>

and the view and capture the source ...


----------------------------------------------------------------------

Subject: RE: Trig problem
From: "Sadowski, Allan" <allan.sadowski@ncshp.org>
Date: Sat, 4 Nov 2000 12:05:19 -0500
X-Message-Number: 10

Some interesting comments... and I too have made some efforts towards
trying to solve the same problem...

My apologies on the length of this response... I've been trying to crack
this nut
for a couple of years... cheaply that is...   it would have serious
implications for 
many applications... if I can cheaply determine position accurately in
multistory
buildings... without having to prewire the building with infrastructure to
support
said effort...  there are a number of vendors that prewire to do this... but
at best... 
they tell you which room you are in (accuracy wise this is fine) .. but you
have to
wire room/building in advance... not an acceptable part of the solution I
need.....

The "problem" that I keep having... is how to accurately and consistently 
do a "time hack" at the same point of the sampled received waveform ... 
without having to use incredibly high bandwidth signals with associated
expensive sampling and processing equipment to get rid of ambiguity...
nearly insignificant "jitter" equates to significant error in the
results....  yes, you
can do some simple statistical processing to get accuracy back... but that
drives the expense and time to get a solution up pretty fast....   heck,
I'd love to consistently get a result within 5 meters....  think of it this
way...
If you had a 100MHz sampling rate.... that means your sample is every
10 nanoseconds.... even if there was no ambiguity ... that still equates to
a
plus or minus 10 ns error in your time hack.... which is approximately 7
meters....

Now.... that approach implicitly assumes you have sufficient post processing
so that you can exactly "hack" (determine which sample is the correct one)
at the right time....   in reality - at those sample rates.... you're going
to have
a heck of a problem figuring out which sample it the one you are going to
reference
from the IR signal.... oh by the way.... you're sampling that one too... so
getting
a consistent "hack" on that sampled signal so you can start your timer
looking
to hack on the ultrasound signal... is also a problem....  repeatability is
key....

Maybe I'm lookin at the problem wrong...  If I had an ultrasound transceiver
on
board the unit.... (and assume I know the dimensions of the room a-priori
and
that there is nothing to confuse the receipt of the signals (obsticles)...;)
then
I'd just put the ultrasound gear on a turntable... and get distance samples
in all
directions.... and from the distances... I'd be able to pretty accurately
determine
where I was in the room... since I have the room dimensions already....

Anyway... the problem I have... is where in a multistory building I am...
when I cannot
wire the building in advance.... and obviously... GPS doesn't go through
many 
buildings very well....   Heck... I'd love to just  change a gps to a
different freq... and
set up around the building ... surrogate GPS "satellites".... all with
enough power
to penetrate the building... and then use the well proven signal processing
that
GPS receivers already have...  ah... and a bullfrog wishes it had wings....

Simplistically.... back to the issue... if you were doing a simple energy
threshhold 
function.... it's not a true step function.... it's too has a finite rising
edge time.... 
so how do you get your instruments to "hack" at the right time.... even if
you 
increase your sampling rate to 200 or 500MHz.... the amount of slop (error)
will 
probably exceed the size of the room.... (given the above relative
difference of 
arrival approach mentioned at the start)...

A neat problem.... and I'd certainly love to work on a solution....

Two approachs I looked at were:

Conceptually, have a 802.11 card transmit x pings a second.....  hack some 
802.11 chip sets (by the way, this is the 11Mbps wireless network
technology) 
do a time hack when the chip sequence was an exact match.... (it's a sliding

window... when the convolution is 1 or -1 it's an exact match).... the error
function 
(time ambiguity) here is still on the order of 20 to 40 nanoseconds - I
think...

Now... the next part of that same problem is that I need high accuracy
position
info for each receiver... given survey techniques... we can get 10-30 cm
with the
existing technology.... relatively fast.....    So.... after each receiver
gets an 
exact time "hack" for when it receives the beacon (ping).... and having the 
"exact" position of each receiver.... we can do some simple math and solve
where the beacon is in threespace....

I decided the cost was a killer with the above approach.... so I came up
with a
cheaper concept... but it too has problems... bandwidth is one.... and the
accurate
unambiguious time hacks...

The "beacon" is a tiny low power full-dup repeater.... ... a master station
sends out a
"waveform" that the beacon just repeats.... simple... there's more... there
are a
number of accurately positioned beacon receivers.... just like mentioned
above...
these beacon receivers also do an accurate "time hack" on the received
signal (we'll
overlook the obvious issue of how you get an accurate "time hack"... it's
been 
addressed above...) so each beacon receiver sends the "absolute" time hack
back
to the master... and since it knows it's own position... the position of the
receivers,
and the time delays involved.... again a simple bit of math... and the
"repeater" beacon's
position can be determined.....    problem here (in addition to bandwidth
needs for
the "waveform" - in order to get decent position accuracy).... was trying to
get absolute
time - consistently across multiple beacon receivers.... no good....

SO... the next thought was... heck... just eat up more bandwidth... and have
each
beacon receiver actually be a full dup repeater too... to save some
bandwidth... each
beacon receiver would be alotted it's own time... instead of all having
their own second
freq... and the multitude of receivers at the master station is reduced to
one (or two if it's 
also listening to the beacon itself)...  anyway.... this gets rid of the
problem of having to 
sync all the beacon receivers to the same clock... so it really is a simpler
approach... 
but still... how to get an unambiguious consistent time hack....

At this point... maybe I just need to build it... and see what happens !!!!

ALOHA
AH6LS

> ----------
> From: 	Wes Johnston[SMTP:kd4rdb@usa.net]
> 



Read previous mail | Read next mail


 17.11.2026 03:13:52lGo back Go up