Reconstruction of WSPR Signals for Opportunistic Sensing

[Tejas Bhagawatula]

Hello everyone,

I'm a high school student (and member of the PSWS curricular integration group) working on a project inspired by HamSCI's propagation research, and I'd appreciate any feedback.

A significant amount of citizen science-based and opportunistic propagation research has focused solely on decoder outputs, such as SNR, spot count, and Doppler shift.

I'm interested in understanding whether this may obscure short-term and/or waveform-level consequences of atmospheric processes, both in the ionosphere and the troposphere.

Since modes like WSPR are reproducible and publicly available, they allow us to compare the reconstructed ideal transmitted waveform versus the received signal, leading to potential for waveform-level analysis of channel effects.

In essence, I'd like to find out whether the residual contains propagation observables not present in conventional WSPR spots.

The project is still in its early stages, but I've written a more detailed technical description at this GitHub repo.

I'd really appreciate feedback on whether this direction seems scientifically useful and whether there are existing datasets, previous work, or researchers I should be aware of.

Thank you very much for your time, and I'd love to hear any thoughts or suggestions!

73,

Tejas Bhagawatula

KJ5OAE

[Philip Gladstone]

I will note that you can do the same with FT4/FT8 signals. I've done some work on matching the entire received signal against the predicted waveform and it turns out to be difficult to get the exact frequency and start time of the signal. 

The plot below is the correlation magnitude of a particular FT8 signal against (audio frequency) and time. As you can see, the frequency needs to be pretty spot on to get anything, but the time can be within 10mS or so. The X marked here is the correlation peak. 

Good luck

Philip

p.s. My work is in trying to perform TDOA analysis of FT8 transmissions.

Philip Gladstone

>

--
Please follow the HamSCI Community Participation Guidelines at http://hamsci.org/hamsci-community-participation-guidelines.
---
You received this message because you are subscribed to the Google Groups "HamSCI" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hamsci+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/hamsci/1ed86866-1b99-4571-9405-0c7a36add3bdn%40googlegroups.com.

[Tejas Bhagawatula]

This is very interesting work --

I was planning to match a cross-ambiguity function to mitigate the phenomena you're describing, which (in my understanding) views the signals on a similar plane that the figure you attached does, and offsets the received signal by the differences in time and frequency dictated by the peak magnitude of the graph.

I haven't gotten to testing this, however, so I'm not sure how well it would work/how practical it would be yet.

I'm excited to hear more about your FT8/4 TDOA findings!

Thanks,

Tejas Bhagawatula

KJ5OAE

[Dr. Nathaniel A. Frissell Ph.D.]

Hi Tejas and Philip,

This is very interesting work that you are doing here. I think it would especially be of interest to our HF Channel Sounding working group. Some people particularly interested in this would be Bob Johnk WB0VGD, Kristina KD8OXT, Phil W1PJE, Ethan K8GU, and Bill NQ6Z (CCd).

I’d like to figure out how to better support you in this work. Send me an email and we can find a time to talk more about it.

73 de Nathaniel W2NAF

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

Nathaniel A. Frissell, Ph.D., W2NAF

Associate Professor

Department of Physics and Engineering

University of Scranton

HamSCI Lead

From: ham...@googlegroups.com <ham...@googlegroups.com> on behalf of Tejas Bhagawatula <tejas.bh...@gmail.com>
Date: Sunday, July 19, 2026 at 10:43 PM
To: HamSCI <ham...@googlegroups.com>
Subject: Re: [HamSCI] Reconstruction of WSPR Signals for Opportunistic Sensing

WARNING: This message originated from a source outside of the university.

To view this discussion visit https://groups.google.com/d/msgid/hamsci/f39ce51d-b6f1-4c2f-b511-15a4fa69238dn%40googlegroups.com.

Please use your best judgment when opening attachments, clicking links, or replying to this email. If you suspect this is a phishing attempt, please forward to inf...@scranton.edu. Other questions can be directed to the Technology Support Center at (570) 941-4357 or techs...@scranton.edu.

[Phil Erickson]

Hi all,

  Philip mentioned this to me at a meeting a couple months ago.  I am particularly interested in the statistical variability of the correlation peak as a function of time by binning together various spatial regions.  Also interested in whether the correlation peak and variability is symmetric or asymmetric with respect to the sender.  One has an advantage there in that FT4 and FT8 are two-way communications for successful QSOs so you are guaranteed to probe both directions.

73

Phil W1PJE

-----

Philip Erickson (he/him), Ph.D.

Director

MIT Haystack Observatory

Westford, MA  01886  USA

email: p...@mit.edu  [NOTE NEW PREFERRED ADDRESS]

WWW: http://www.haystack.mit.edu

voice: +1 617 715 5769

fax:   +1 617 715 5590

Public key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x54878872

[Rob Robinett]

One problem with using FT4/8 top reports is that pskreporter can filter out 90+% of the spot reports. 

It saves only one tx/rx pair every 20 minutes, so there will be only one such record for each direction of most QSOs

Sigmond sites will save every spot to the psk.spot database at wsprdaemon.org, which will soon dwarf the wspr spot database.

To view this discussion visit https://groups.google.com/d/msgid/hamsci/CAAEMTOUWZZ93zd%3DVzuoaikoe-9D-AWBUOkOysi%2BKFMhRopR9mA%40mail.gmail.com.

--

Rob Robinett

AI6VN

r...@robinett.us

mobile: +1 650 218 8896

[Dr. Nathaniel A. Frissell Ph.D.]

I think this is one reason it would be good to for Tejas to talk directly with me and the HF Channel sounding group. To truly effectively do what he wants to do, we may need to set up PSWS nodes to run certain correlation/custom algorithms right on the nodes themselves and then report the data back to the central servers. I think this is a good example of how we might be able to expand the usefulness of our system without simply recording large amounts of IQ data.

73 de Nathaniel W2NAF

[Philip Gladstone]

The PSKReporter team is working on increasing spot storage so that we can save more data. I have a version of the ft8/ft4 decoder from wsprdaemon that can compute the time/frequency of each transmission and is still able to run on a mini-pc (assuming that only a < 5kHz frequency band is being analyzed for each band). The code is rather dirty! 

However, I can compute the rough location of the transmitter based on the TDOA data. [I'm using a bunch of kiwisdr receivers to gather lots of reports of the same transmission.] 

I'll put up a patch set so that other people can look at it.

Philip

Philip Gladstone

To view this discussion visit https://groups.google.com/d/msgid/hamsci/BN7PR03MB4403C32E0C9BD10E83EB4E64F2C32%40BN7PR03MB4403.namprd03.prod.outlook.com.

[J. Gordon Beattie, Jr.]

Nathaniel,

Targeted embedded functionality in a device based on specific chronic or persistent requirements or objectives makes sense.  We could identify specific event signatures or fingerprints that qualify a specific node and do comprehensive management of that node or neighbors to provide improved insights and issue resolution is something I started doing at a prior employer as a way to effectively manage scale and scope for improving whole groups of customers.

Maybe instead of "drinking the ocean", we cache collected data and only push enhanced functionality down to the probe or collect data from the probes or a subset of them when needed.

Thoughts?

73,

Gordon Beattie, W2TTT 

201.314.6964 

W2...@ATT.NET

Gordon.B...@VIAVISolutions.us

Get On The Air!

---

From: 'Dr. Nathaniel A. Frissell Ph.D.' via HamSCI <ham...@googlegroups.com>
Sent: Monday, 20 July 2026 09:27:44
To: Rob Robinett <r...@robinett.us>; ham...@googlegroups.com <ham...@googlegroups.com>
Cc: Robert Johnk <robert...@gmail.com>; Ethan Miller K8GU <et...@k8gu.com>; Bill Liles <lil...@gmail.com>; K. Collins <kv...@case.edu>

To view this discussion visit https://groups.google.com/d/msgid/hamsci/BN7PR03MB4403C32E0C9BD10E83EB4E64F2C32%40BN7PR03MB4403.namprd03.prod.outlook.com.