« Wow! I got 19 out of 21
Main
An Experiment with a Little Content »
February 3, 2006
Agile in a Clumsy Sort of Way
I don't blog much about the technology I was involved with in midlife but the New Scientist has an article on a software-defined radio test that did catch my normal, if not my abnormal, interest. When I was involved in the start up the wireless group at Rockwell International's Semiconductor Division (now Conexant; well, the wireless group is actually now called Skyworks, Inc.) we spent a lot of our stockholders' money discussing what we called "common air interface agile radios." What this really meant was a very smart (and cheap) radio that could be configured using software to receive and transmit any digital or analog modulation scheme, at any frequency and supporting any communications protocol from very broadband wireless LAN to the several cell phone technologies to cordless phones for any type of communications: voice, data, navigation, control and multimedia.
On top of this would be a high level management software system in the handheld device that selected the wireless service (the common air interface) that was least costly for the requested application. For example, if you wanted to place a voice call, the system would look for the cheapest service available. If you were at home, it might be your cordless phone. If you still had free minutes with your cellular service, it might use them or ask you if you wanted to use them. If you were out of range of your cordless phone base station and your cellular phone service, it would try a satellite-based service. If on the other hand, you wanted to send a picture, it wouldn't even try the cordless phone because it would know that the cordless phone could not support the transmission of a picture. Or is you wanted to send data, the system would use the size of the data file to select the optimum tradeoff between cost and speed of transmission. You could set preferences. Movies and streaming audio (at least high quality audio) would come or go over the best-matched wireless system. And on and on. This technology is sometimes called local least cost routing. The service providers didn't like this idea very much and they still don't.
If all this sounds like wireless nirvana, in some ways, it is. A whole raft of new technologies was needed (and still are needed) to be developed to make all this work. Here's a passage from the New Scientist article.
Although software-defined radio devices use a normal antenna and amplifier to receive a signal they are fundamentally different from conventional radio-based equipment. An analogue-to-digital converter changes the signal into a digital format, which can then be processed and manipulated by the software. And the software can reconfigure itself to let the device retrieve information sent at alternative frequencies or encoded (modulated) in a different way.
What this is telling you is that a software defined radio consists of an analog section (the antenna and amplifier(s), a conversion section (the analogue to digital and digital to analogue converters) and a software controlled digital signal processing section. The less of an analog section, the better.
Well it looks like such a radio, likely without all of the management layer I described above, is about to be field-tested in Ireland by the Centre for Telecommunications Value-Chain Research (CTVR) in Dublin.
Now here's another passage from the article,
The CTVR trial will involve testing communications between software-defined radio devices across the radio frequencies of 2.08 gigahertz to 2.35 GHz, at several sites across Ireland.The researchers will try switching the radios between frequencies and modulations for different applications, such as audio and streaming video or data transmission, and will also let the devices automatically select the best standard to use.
Oops, the total frequency bandwidth is only from 2.08 GHz to 2.35 GHz. This is a good start. But 270 MHz centered around 2.215 just isn't very agile. For example, it would not support GPS services in the 1.227 GHz or 1.575 GHz ranges. And it sure won't support ultra-broadband on the one hand or open your garage door and turn on your house lights on the other hand. And there's a whole lot else now in use or in planning that it won't support either. The military and commercial aviation have radios that could do much better than these software definable radios that the CTVR is testing. But, they are expensive and big and power consumptive but otherwise not new technology. There have also been limited tests of such radios here in the US but not on the scale of the tests to be done in Ireland. The limiting factors are still those analog portions of the radio that are nearly impossible to design so they work over an extremely broad spectrum. In addition, the speed, accuracy, percussion, power consumption and cost of the analog to digital converters and the speed, power consumption and cost for the digital processing engine are also limiting factors. It turns out that the software portion, while not easy, is current state of the art because for every radio system that the software configurable radio would need to run has already been written for a single service engine. It is just a matter of porting it. Of course, that's easy for an old marketing guy to say.
Despite the limitations, it's good to see that those dreams we dreamt over fifteen years ago are beginning to be realized. It's at times like this that I do get a itch for my old career. The good news is that I can scratch it and it goes away.
Posted by Duane Smith at February 3, 2006 8:49 PM | Read more on Science - General |
Trackback Pings
TrackBack URL for this entry:
http://www.telecomtally.com/mt/mt-tb.cgi/1561
Comments
Sorry, comments are closed for this post.
Send me an email if it is important.