Re: [Discuss-gnuradio] Eye Pattern visualization & Doppler-based perigee estimator

[Jim White]

Hi Marcus,

Thanks very much for the reply.  And the hint on topic.

I think the Doppler problem is easier than you describe.  And I'd note 
it's not a perigee estimator it's a TCA calculator as below.

The problem we are trying to solve is that when a rocket launches with 
multiple payloads (two in the past two weeks with about 40 and 64 
respectively) it becomes difficult to figure out which set of orbital 
elements belongs to which satellite.  We all rely on the Air Force Joint 
Satellite Operations Command (JSPOC) to target the satellites with 
various radar and optical sensing resources then create a set of values 
that define the orbit of each satellite (commonly called NASA 2 line 
elements sets, or Keplarian elements, or keps).  When you read those 
into satellite tracking software it shows you where the satellite is and 
when it comes over your location that software points your antennas at 
the satellite.  Many of these satellites look identical to the sensors 
(they are "cubesats") so the JSPOC can't tell them apart.  They start 
out very close together (at most a few seconds different in when they 
come over the horizon) and over a few weeks drift apart.  You need to be 
able to point your antennas at the satellite to receive data from it and 
command it.  Which means you need to use the set of keps that matches 
your satellite.

In the period from a few days after launch to perhaps a month after 
launch, when owners are talking intensively to their satellites to get 
them up and running (commissioning), JSPOC is distributing a list of 
keps, one for each satellite, but no identification of which is which.  
It's up to the owners to figure that out so a namecan attached to each 
set of keps.

The satellites transmit either short bursts of data or longer 
transmissions containing lots of data as they pass over.  The radio 
frequency changes during a pass because of the Doppler effect.  For a 
satellite transmitting on UHF (about 440 MHz) the signal as received on 
the ground starts out about 10 Khz above the actual frequency and ends 
about 10 Khz below it.  If you plot frequency vs. time you get a COS 
curve (you can find lots of these by Googling Doppler shift).  Each 
satellite is coming over at a slightly different time.  So the frequency 
vs. time plot for each will be different.

If you curve fit the plot of points (or nearly continuous signal) you 
get a COS curve.  The point at which the curve is steepest is the point 
in time when the satellite was closest to your station (TCA, Time of 
Closest Approach).  If you put the keps from all the satellites that 
were launched together - any one of which may be yours - into a tracking 
program that program will calculate (and show on real time world map) 
the TCA for each.  (see http://gpredict.oz9aec.net/) So the solution to 
figuring out which set of keps is your satellite can be solved by 
matching your observed TCA with the calculated TCA for each set of keps.

As noted, figuring that out in that first month or so is critical 
because satellites need some care and feeding initially to get them 
pointed at the sun for battery charging, get them into their science 
data collecting mode, etc.  And that time period is when everyone is 
scrambling to figure out which set of keps is for their satellite so 
they can point their antennas properly.

The calculation / math problem can I think be tackled in two steps.  The 
first is to curve fit a COS equation to the recorded set of time vs. 
frequency points during or after a pass.  The second step it to find the 
point on that curve where the slope is highest (most up and down) which 
is the time of TCA.  One could then manually/visually match that time to 
the TCA calculated/displayed by the tracking program for each set of keps.

Many folks are using SDRs to talk to their satellites now and most are 
using GNURadio with the SDR (in the current case we are using an Ettus 
B200).  Our GNURadio flow graph provides a spectrum analyzer display 
where we can see the signal along with a real time output to a telemetry 
decoder program which shows what's happening inside the satellite.

The first step would seem to be a block that plots in real time 
frequency vs. time of each burst of data from the satellite. The display 
would be a rectangular X/Y plot that would look much like the Doppler 
curve plots you can Google for.  That could then be advanced by doing a 
real time curve fit to those points and overlaying it on the plot.  And 
the last step would be to calculate in real time (or after the pass) the 
steepest point on that curve.

One might wonder how often this would be useful.  5 years ago I would 
have answered about once every couple of years.  But the number of 
launches of flocks of tiny satellites is growing exponentially as 
evidenced by three launches of a combined 100+ satellites in the past 
three months.  That pace won't keep up but 4 or 5 multiple payload 
launches a year is likely.  And in some cases solving this problem for 
the owners can mean the success or failure of their satellite project, 
many of which are built by universities and even high schools.

I hope that's a better explanation of the problem and how it might be 
solved.  I had hoped existing code/blocks would be available that might 
be combined to implement this function. Perhaps with this explanation 
something might be found that will work.

Thanks again,

Jim

On 12/5/2018 4:41 AM, Müller, Marcus (CEL) wrote:
> Hey Jim,
>
> welcome to the community!
> A quick hint beforehand: When you change the subject line (as the mail
> you responded to suggested strongly), then people can more easily guess
> whether they should be taking the time to read your mail. I was about
> to skip it, and that would've been a bummer!
>
> So, an eye pattern display is currently not available in GNU Radio
> itself. We "only" have the QT GUI Time Sink, but that has no
> "persistence" setting as of now. It's on our list of things to
> implement for halfway feature-completeness. I'm optimistic we can come
> up with a solution to that together. Can you talk a bit about your
> specific use case? Is your GMSK reception already synchronized?
>
> Regarding TCA: hm, I'm not overly familiar with the math involved, but
> sounds like you'd observe the Doppler development over time to estimate
> the slope of a set of tangents to the elliptical orbit of the satellite
> and then solve for the time where orbit intersects with the main axis.
>
> So: a bit of a math problem :) which could be relatively nicely solved
> by first implementing a Doppler estimator (how definitely depends on
> your system's characteristics and speed needs), and using the velocity
> data that spits out either in post-processing of recorded speeds, or in
> a python GNU Radio block that accumulates these velocity estimates and
> estimates a TCA based on that.
>
> Best regards,
> Marcus
>
> On Tue, 2018-12-04 at 10:15 -0700, Jim White wrote:
> > Very new to GNURadio here...
> >
> > Is there a block available that will display an eye pattern for
> > digital
> > modulation for a GMSK or FM receiver?  I'm using a USRP B200 and
> > have
> > built a couple of simple flows for it including a spectrum
> > analyzer.
> > I'm looking for something that will show the eye pattern.
> >
> > Also hoping for a block that will calculate the time of closest
> > approach
> > from the Doppler curve after listening to and recording a satellite
> > pass.  This is about finding the TCA then matching it to calculated
> > TCAs
> > provided by tracking programs. The objective is to match observed TCA
> > to
> > TCAs from several sets of keps to match them up and determine which
> > set
> > of keps is the satellite we are working with.
> >
> > Jim
> >
> >
> > On 12/4/2018 10:00 AM, address@hidden wrote:
> > > Send Discuss-gnuradio mailing list submissions to
> > >         address@hidden
> > >
> > > To subscribe or unsubscribe via the World Wide Web, visit
> > >         https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> > > or, via email, send a message with subject or body 'help' to
> > >         address@hidden
> > >
> > > You can reach the person managing the list at
> > >         address@hidden
> > >
> > > When replying, please edit your Subject line so it is more specific
> > > than "Re: Contents of Discuss-gnuradio digest..."
> > >
> > >
> > > Today's Topics:
> > >
> > >      1. Fwd: [hamsci-swstation] A Little Fun with Digital RF
> > >         (Michelle Thompson)
> > >      2. Re: SDR for beginners (if not quite "dummies") (Bruce
> > > Ferrell)
> > >      3. Call for Participation FOSDEM & pre-FOSDEM Hackfest (Andrej
> > > Rode)
> > >      4. Looking for other users of the XTRX SDR (Markus Wirsing)
> > >      5. Re: Looking for other users of the XTRX SDR (Matthias
> > > Br?ndli)
> > >      6. Re: Looking for other users of the XTRX SDR (Markus Wirsing)
> > >
> > >
> > > -----------------------------------------------------------------
> > > -----
> > >
> > > Message: 1
> > > Date: Mon, 3 Dec 2018 16:05:20 -0800
> > > From: Michelle Thompson <address@hidden>
> > > To: GNURadio Discussion List <address@hidden>
> > > Subject: [Discuss-gnuradio] Fwd: [hamsci-swstation] A Little Fun
> > > with
> > >         Digital RF
> > > Message-ID:
> > >         <
> > > address@hidden>
> > > Content-Type: text/plain; charset="utf-8"
> > >
> > > I enjoyed reading this writeup from a GNU Radio user, and wanted to
> > > share
> > > it.
> > >
> > > -Michelle W5NYV
> > >
> > >
> > >
> > >
> > > ---------- Forwarded message ---------
> > > From: Greg KF5N <address@hidden>
> > > Date: Sun, Dec 2, 2018 at 5:01 PM
> > > Subject: [hamsci-swstation] A Little Fun with Digital RF
> > > To: hamsci-swstation <address@hidden>
> > >
> > >
> > > I had a great time at the TAPR Conference in Albuquerque, and I was
> > > especially intrigued by the Space Weather discussion in the Sunday
> > > seminar.
> > >
> > > Nathaniel's description of "Digital RF" really got my attention,
> > > and I
> > > decided I wanted to experiment with it.
> > > That took some doing, because I had to get a receiver and come up
> > > to speed
> > > on GNU Radio.
> > >
> > > So about 3 months later playing with radios, computers, and
> > > software here
> > > is a report on my experiments:
> > >
> > > https://github.com/Greg-R/explore_digital_rf/blob/master/docs/explore_digital_rf.pdf
> > >
> > > The top of the repository is here:
> > >
> > > https://github.com/Greg-R/explore_digital_rf
> > >
> > > I want to do more experimenting using HF band
> > > signals.  Unfortunately I am
> > > in a hostile RF environment.
> > > I've got my hands on a broadband loop antenna, so getting that
> > > mounting and
> > > working is next.
> > > I hope I can pursue further experiments with the combination of GNU
> > > Radio
> > > and Digital RF.
> > > Cool stuff!
> > >
> > > 73 Greg KF5N
> > _______________________________________________
> > Discuss-gnuradio mailing list
> > address@hidden
> > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio