Linux Powers First Handheld Software Radio

An anonymous reader writes "According to this article at LinuxDevices.com, Vanu Technology is demonstrating what it claims represents the world's first handheld 'software radio' using an iPAQ PDA running Linux at a conference in Washington DC today. Vanu apparently has implemented the signal processing functions on the iPAQ's XScale processor, and their software uses POSIX APIs to make it platform independent. Software radios implement multiple radio standards and frequency bands in software, rather than hardware. A standard iPAQ expansion pack houses the radio transceiver."

Comments

One of the concerns

[Froze]

of software radio is the ability to modify the code and tromp all over someone elses legally protected frequency range. Some of the big nonos include sending on ATC (air traffic controll) frequencies and numerous other military and civil service bands.

Someone corect me if I am wrong, but couldn't the transceiver be built with hardware filters on those bands and thus sidestep the issue of broadcast interference? I know this is not as nice as having a fully programmable software radio transmitter, but otherwise I really don't see the FCC granting any kind of production licensing for these.

Anyone else have solutions to this dilemma?

Is it a real "plus" ?

[chrispy666]

Unless the software solution offers significant improvements/advantages (like super fine tuning, rms, ta, like the stuff on the car audio tuners) and catching international radio stations, I just don't really see the point of having linux in something that works great as it is already (i.e., the old fashioned way)...

Re:Call me old fashioned...

[evilviper]

One of the benefits of software defined radio is that you can tune in any number of channels simultaneously...

This would have many uses in the high-end radio communication field. Although I'm not sure anyone really needs a handheld version...

CF version?

[Derg]

Is there the possibility of turning what currently fits into a iPaq addon into a CF card? That, imho, would be a really killer device..

just a poor geeks dream...

Expansion pack

[Stonent1]

Isn't it just a pcmcia interface? So potentially this item could be reproduced for a laptop computer?

Yup :-)

[sonamchauhan]

Funny you say that... http://www.linradio.com/

This is a software-defined-radio PCI card.

next privacy issue?

[torre]

What exists to stop this from becoming the next supper scanner?...

The current configuration of the device is said to support commercial analog FM radio services, including Family Band Radio as well as the public safety APCO 25 digital standard, with future prototypes under development that will include operational capabilities of up to 900 MHz and support for cellular and PCS standards such as TDMA and GSM.

With such a huge frequency range under its belt and the fact that it's all process via software all it needs is some voice recognition software and it could become the ultimate scanner/big brother toy. Simply put, you enter a few key words, and it scans the airways for you looking for them until it finds them and either logs it or tunes you into it. The NSA has had stuff like this for listening in on international call, but I don't know if I like the idea of my neighbour being able to selectively listen in on my calls especially with such power...

me->Hi I'd like to buy blah
staff-> will that be Visa or MasterCard
me-> Visa...
person with smart scanner->Chaching!

Re:next privacy issue?

[torre]

"Dude I can't get voice recognition to work reliably when fed by a voice cancelling headphone on a 1.2Ghz machine and a couple hours of training,..."

Funny.... I've had reliable keyword recognition since way back in '95 via my P133 and IBM via voice on OS/2 and the Mwave dsp addon board. Keyword isn't that big of a deal (relatively that is) if your selected vocabulary is relatively small. Keyword voice recognition is an almost solved problem and is used often from automated phone systems with amazing accuracy given poor signal quality. The automated collect call system's come to mind as a simple example. I have seen more complicated systems in work which are currently in research and some of the toolkits are open sourced if I'm not mistaken (would have to check to be sure)

As for difficulty.... that's not difficult at all seeing that the goal of the project is to ultimately provide that functionality.

from they're website
"IS-136, IS-95, GSM
A complete cell phone implementation"

All some interested party has to do is take they're freely available cell phone friendly code when it comes up make some modifications so that the signal is piped through a keyword recognizer instead of the speakers and poof... Its not as hard as you claim when sooo many people are willing to give you what you need.

"It would cost a lot more than $14K to do this over 100 channels."

As for cost, first of all I was using fictitious numbers as I stated... but seeing that you've brought up cost as unrealistic... I'll bring more realistic number and now overestimate. First, the paper associated with the article states that a dual 2Ghz machine could handle upto 32 GMS channels.. so.. lets see... lets say it costs $100 (which is probably being conservative) in Radio shack hardware to make the hardware to support 1 channel seeing that they say it only requires fairly inexpensive hardware... that's 3200 for 32 channels. Now add an overpriced Dell dual CPU server @ 3,444. now to get 96 channels it would cost you $19,932. Now, the average person could dig up a dual machine for less than 2000 with similar spec (minus scsi)... so realistically.... It be more like $15,600 which isn't that far off from my original $14k fictitious guess.

"Now targeted scanning could be a problem, but then maybe I can get my freaking cellular provider to turn on basic GSM encryption (phone supports it but none of the cell sites in the US do AFAIK)."

Unfortunately encryption won't end up being much of a stopping stone. It however will probably be the hardest part to deal with. There's a couple of ways that could be thwarted. 1. with so many distributed system for encryption breaking, a brute force could be used, not elegant... but hey... it works... 2. social engineering could be used to gain access to the information... Its certainly not the first time that a provider's phone password has popped up on the net. 3. if the original purpose of the system trickles down to the cellphone/pda... then you could simply walk up to your service provider and ask them to set it up, and voila you know have the encryption key.

I could keep circumventing different ideas all night.... And that's what's scary... cell phone scanners currently exist but are really pricey. Adding the cost of one sc

Re:Call me old fashioned...

[GoRK]

Erm.. No. That's not technically true.

Although one of the little whiz-bang demos of software radio generally involves tuning two FM radio stations or something at the same time, there's nothing particularly unique about a software-defined radio that makes it possible. Couple the right wideband receiver with the right circuts to do some off-center modulation and you could build an analog radio that would tune a couple stations at the same time too.

It's simply a question of how much bandwidth you can tune simeotaneously, how much bandwidth each component signal occupies, and the wideband rx having enough definition to clearly modulate a signal that it is not directly tuned to receive.

Some of these issues are really going to be stickers with bringing software-defined radios to the general market on a large scale. Yes, in theory a software defined radio might be able to tune AM, FM, HDTV, 802.11b and every cellular protocol ever, but actually producing the analog part that could do the RF job for that software radio would be a real trick indeed! The radio in this article, for instance, only does 100-400MHz or so.

Let's get some good software-defined antennas going here (Phased Gate Array antennas have some good promise here), and perhaps some software-defined RF electronics (think the analog equivalent of an FPGA) and then we'll really be in business for this software-defined junk!

~GoRK

SDR has been done, but not this small

[mrFur]

SDR has been done before. WinRadio has a "DC to Daylight" receiver that fits on a PCI card, and allows you to play with the demodulation. A lot of amateurs are working are working with this, and one gives you the VB source for the demodulator with a pretty front end. (I don't want to /. him, so look for SDR-1000 at Google).

The interesting part of this is that it was built into a hand-held computer for the first time, and the practical implementation means that any new radio service is a software upgrade. Think back to the Telco's when call display came out. Instead of upgrading the phones, they had to upgrade the entire network just for a service. This is an infrastructure for radio like the Internet, where services are layered over the basic medium.

Re:does it do DAB?

[BigBadBri]

It doesn't look like it from the specs, and with 22% CPU utilisation just for FM reception, I doubt that a PDA has the guts to perform the DAB decoding in anything like real time.

Besides, £100 gets you a pretty capable DAB receiver here in the UK (if you're lucky enough to find one in stock), and I'd guess that the add-on card for the PDA costs at least that much.