When You Own a UWB Hammer, Even Low Data Rate Problems Look Like High Data Rate Nails

This just in from Information Week (via Dwayne Hendricks, WA8DZP on Twitter): Ultrawideband Takes Us Closer To Star Trek Tricorders. Is it a miraculous UWB radar that monitors your heart rate, blood pressure, and cholesterol? Alas, no. It’s a UWB body area network.

Ultra-wideband works great when you have lots of data to send in a reasonably short line-of-sight link. Want to dump an hour’s worth of video a few feet from your camera to a computer in a matter of seconds? Or send a couple high-definition video channels to your TV? UWB could be the solution for you. But if I understand this body-area network [BAN] business, we’re talking low data rate sensors. From the article:

[Body area networks] which are computing devices worn on, in, and near the body for health monitoring purposes, are generating increased interest among the medical and high tech community as the technology can capture physiological electrical signals from the human body including brain waves, heart health, and muscle response.

According to the good folks at Analog Devices (who make it their business to figure things like this out for their customers), “[a] standard clinical ECG application has a bandwidth of 0.05 Hz to 100 Hz.” Their Application Data Sheet shows a 12bit ADC. Let’s do the math. 12bits of data at 100Hz is 1200 b/s. Let’s assume any error correction overhead and data compression balance each other out. We’re talking just over 1kb/s per channel.  How many channels do you really need in your BAN? This sure sounds like a low-data-rate application to me. Why wouldn’t a 100kb/s link work just fine for this with capacity to spare?

The other problem with UWB for a body area network is with the propagation characteristics of the short wavelengths. They tend to behave in a quasi-optical fashion. UWB signals will be blocked by the body or other obstructions. And let’s hope you don’t roll over in your sleep on top of your UWB telemetry radio. That’s exactly the kind of problem the researchers found:

Through their experiments, the OSU researchers found that ultrawideband might have that capability if the receiver getting the data were within a line of sight and not interrupted by passing through a human body. But even non-line of sight transmission might be possible using ultrawideband if lower transmission rates were used.

Certainly, a UWB link can be designed to degrade gracefully over wildly varying channel characteristics. That’s a clever piece of system engineering and I respect anyone who can pull it off successfully. But if the system is still feasible with “lower transmission rates” from a non-line-of-sight UWB link, why use UWB in the first place? A decent VHF or even lower frequency telemetry data link ought to work better in this application. I’d be looking at modifying a near-field communication link operating at 13.56MHz to extend the range from 4-20cm out to 3-5m to create a BAN that wouldn’t care if you rolled over on top of the transmitter.

By way of comparison, here’s an application of UWB data links in medical systems that makes some sense (Hat tip: Steven J. Crowley). A couple of weeks ago, Ueda Japan Radio Co Ltd demonstrated a 480Mb/s UWB data link for connecting diagnostic equipment (that looks like ultrasound) with printers and visual displays. There doesn’t appear to be much application-specific system integration. It looks like they took someone’s off-the-shelf UWB data link and hooked it up between an ultrasound machine and a control computer. But still, this is an example of a “medical application” where UWB has value.

UWB is an exciting technology. It can be made to fit in a wide variety of applications. But it is not necessarily the best solution for all problems.

Leave a comment

Your email address will not be published. Required fields are marked *