RTLSDR Collaborator wanted.
Douglas Miller wrote 04/01/2019 at 17:14 • 0 pointsI need someone to collaborate on a RTLSDR project to 'give something back' to first responders in my area. It might even save lives. What's needed is a GNURadio or (preferably) a DLL that can decode pi/4 DQPSK packets.
Working with SDR's is an area I know very little about, so I'm asking for help. All the rest of the programming needed I can do, I just need the demodulator for the packets. If anyone's interested in
taking it on, please PM me and I'll explain the rest.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.
Here's what I'm trying to get done.
The city where I live has three major train tracks pretty much cutting the city in two, with over a hundred
freight trains a day running through it. The tracks are only a block from the center of the city. The intersections are frequently blocked, sometimes for hours and now and then for more than a day as
the trains wait to get into the freight yard a few miles away. It's one of the top five yards in the country.
Hackaday did a post on my attempt to see exactly how many trains go through here: https://hackaday.com/2017/11/06/trainspotting-with-junk-for-science/
You can imagine what all those trains do to emergency vehicles. I have a scanner, and it's a daily thing to
hear 'I'm railroaded again' on it, meaning it's stuck waiting for a train. I've personally seen many times an ambulance, police or fire truck on a call sitting at the crossing close to my house with red lights flashing, waiting to cross. Btw, I've been in one of those ambulances, headed for the hospital when I had a heart attack, so it's rather personal for me.
So my idea is to give the first responders a heads up when they are on a call. A web site, along with
a phone app, that shows them a real time map display of where the intersections are open and which are closed or about to be closed, giving them a chance to pick a better route. Maybe a display at the dispatch center, so they can steer everyone around problem areas.
The city is aware of what I'm trying to do. The chiefs of both fire and police would like to see it, and one
city councilwoman is pulling for it to get done. But they don't have the budget to really help out. My plan is not to charge a penny for anyone to use the final app, which of course means I don't have a way
to pay for a pro at this kind of thing. This is a 'give back' to first responders, not a make money, project. I guess the good news is that once it's done it could be implemented in any other city that has a need for it.
I have developed and tested a weatherproof node based on one of Adafruits feather M0's with LORA
radio and a accelerometer/magnetometer. Placed in the vicinity of the tracks, it will radio back a trains location as it passes by when it senses the ground shaking. But it has several drawbacks. One is cost
for the entire system. Another is it won't spot a train that stops dead and just sits there for hours. Plus having to get permission to place one close to as many crossings as I can. Refreshing batteries won't be fun, either, and adding solar panels will sure add to the final cost.
So I had an idea to use the signals already being broadcast over the air. It's similar to the Automatic
Train Control Signals being received by most of the 30,000 members of the ATCS Yahoo Train watchers group. But the signals are modulated using pi/4 DQPSK instead of FSK. There is no guaranty that the
railroad is actually broadcasting the type of packet I need even though I see one type of packet or another coming in several times per second in my area, and there's only one way to find out.
I've worked for months trying to get it done, but, frankly, I'm over 60 and disabled, and learning new things like this is a lot harder than it used to be. I can program, I've been doing it since the late '70s, but wrapping my head around the whole digital signal processing is kicking my ass. I've been told I should expect a nine to twelve month learning period before I get it, starting from scratch.
I know the message protocol in detail, as it's a published spec. Pi/4 DQPSK at either 16kbps or 32kbps
(there are two packet types), and the preamble in hex is “f337eeb637660672”. So once a packet hits my code, I'm golden. But I'm getting no-where fast, hence my call for help.
I looked at the code to decode TETRA, but I'm not connecting enough dots to get it modified for this use.
My total lack of experience in signal analysis is showing. But then I thought of all the Hackaday people who like a challenge and love to help each other out, so here I am, hat in hand. Heck, Hackaday is one
of my first web stops when I get up in the morning.
Anyone willing to help, please email me at Dougmsbbs@hotmail.com.
I'm certainly willing to add your name to the project as a co-author.
And having a few hundred police, paramedics and firemen grateful for your effort sure can't hurt your karma, right?
Are you sure? yes | no
Are you looking to decode TETRA signals? Telive should be able to help you (and I think they have a pi/4 DQPSK demod block for GNURadio). If you're trying to decode voice packets from public services, you might be out of luck, as most of those networks are TEA2 encrypted.
Are you sure? yes | no
I'm not doing TETRA. Please see the new post I just added. It explains what I need to get done.
Are you sure? yes | no
Consider joining #SDR Chat and #RF Chat , you may find people there that are able to help.
Are you sure? yes | no
I will certainly do that. Thanks!
Are you sure? yes | no