Exploring Open-Source Digital Mobile Radio Project

[turbovectorz]

Hi folks!

I am in the early planning stages of building an Open-Source Digital Mobile Radio (DMR) like MotoTRBO following the ETSI specs.
The system will be designed using Software Defined Radio (SDR) test platforms, such as the USRP and SSRP.
The frequency band will be virtually unlimited -- I think we should just concentrate on the White Space spectrum right now which is around 47-850 MHz.

For those who does not have money to spare to purchase the SDR's, I will provide two-way radio GUI emulators running in QEMU or VMWARE. This way we can concentrate on two-way radio features instead of radio propagation issues. Development platform will be using the freely avaliable GCC tool-chain and Eclipse IDE.

Base on the specs, the modulation scheme will use C4FM. I had Sabers using C4FM over the years and they are very robust. However, let me know if anyone has a better idea for modulation such as for P25 adaptation, i.e. DPQSK, CPQSK, etc.

Right off the bat, I will design in the trunking protocols and quite a few vocoders options including frequency-domain (i.e. AMBE/IMBE-like) and time-domain (i.e. MELP, VSELP, GSM HR). Eventually, encryption modules also for much better privacy.

Let me know what your thoughts are!

Sincerely,

TVZ

[MattSR]

Sure, sign me up as a tester. I have a USRP and I know GNUradio well.

Whats the address for the repo? do you have any Alpha code available?

[Wowbagger]

Well, for a vocoder, you might check out this man's work on an improved vocoder.

For a physical layer, you may want to consider using OFDM rather than C4FM - you should be able to get a higher bit rate for the same baud rate (actually, a higher bit rate for a MUCH lower baud rate) and have a more robust transmission layer.

[turbovectorz]

MattSR,

That's GREAT to hear that you have a USRP as well -- Do you need TVRO daughter boards? I am actually waiting for the DC-1GHz boards from Ettus Research (not sure if they are available yet). On another matter, I currently do not have any alpha code yet to release via SVN -- Hopefully, in the next few weeks. Also, I need to finish laying out the draft specs document of which I will share with you folks soon! Overall, we need all the testers we can get!

Wowbagger,

Thank you for the recommendation. I had worked with OFDM in the past (particularly COFDM for DVB-T) in remote surveillance projects operating at very long distances, and it is rather secure.
Also, do you have a link on the vocoder to consider?

You folks are AWESOME keep it coming for feedback/suggestions!

Sincerely,

TVZ

[wavetar]

Also, do you have a link on the vocoder to consider?

It was sort of 'hidden' in the line of his post...you had to put your mouse over the text to see the hyperlink:

http://www.rowetel.com/blog/?p=128

[turbovectorz]

Thank you, wavetar!

Going through the SPR vocoder specs proposed in the blog, it reminds me of a Russian Company that developed a similar technique using TI DSP's:

DSP Innovations Inc. (dispini.com)
Slavi 4, Penza 440000
Russian Federation

David Rowe's work may potentially be a viable option. However, I think we will leave the codec selection as open as possible (i.e. plug-ins) because every situation can be different. In addition, this approach will make it easier for developers integrate software-based voice scramblers and/or automatic noise reduction algorithms.

Sincerely,

TVZ

[MattSR]

Any updates on this? I now have an Ettus WBX transceiver to work with..

Just need some source code to start playing with now..

[turbovectorz]

Any updates on this? I now have an Ettus WBX transceiver to work with..

Just need some source code to start playing with now..

Sorry for the delay folks, so many things surfaced unexpectedly. However, things are finally moving along.
Purchasing USRP WBX repeater boards soon to experiment with and building an alternative small SDR/DSP receiver for the mobiles.

The *half-hacked* OpenDMR simulator is working now, but no source code to release yet.
It's designed using GNU C/C++, gnuradio libs, wxpython and for SDL some really fast waveform graphs/analysis.

Another good reason, is that I also got too involved on playing with Digital Speech Decoder(DSD) lately.

Join us, its fun!


-TVZ