« Modulating information onto radio vortices | Main | MassTLC Innovation Unconference »

October 23, 2011

Comments

Vance Shipley

Great point about MIMO as a game changer. I don't see how you've backed up the statement "just as well at penetrating buildings" though. I looked at the NIST report and it shows a definite decline from 0.5Ghz to 2Ghz and although it's flatter in the 3-8Ghz graphs it's down markedly from the other. Are you suggesting that most of this is the result of dispersion and not absorption? Isn't reflection a large issue here as well which MIMO wouldn't handle?

brough

Thanks Vance. Read section 2.6 on pages 33-35 of the NIST report to see the "post-processing" they apply. Figure 2.6.1 shows actual spectral measurements (before their post processing) for a brick wall. The spectral response is complicated - at one point transmissivity is greater than 1.0! -- but it's relatively independent of frequency. In figure 2.6.2 they show what's happening in the time domain (which they happen to calibrate in meters, based on the speed of light). Figure 2.6.2 shows a primary peak and several additional peaks that result from scattering within the brick, i.e. multi-path propagation.

Figure 2.6.3 shows the post processing NIST applies (in the time domain). This post processing suppresses all delayed, i.e. multi-path, signals. When the post processed signal is then converted back to the spectral domain, suddenly we see what 20th century (non-MIMO) communications receivers see, i.e. RF transmissivity goes down as frequency goes up. However, this is clearly the effect of the post processing. In the raw data, there is no evidence that any of the RF has been converted to heat. It's merely been scattered, thus creating multi-path signals. With MIMO, multi-path becomes a benefit.

As you suggest, MIMO is a game changer. I'm claiming we haven't begun to see how much of a game changer MIMO will be!

brough

Vance, I haven't found good data to show exactly what the scattering looks like, i.e. how much of the signal is reflected back towards the transmitter and how much is merely scattered within the building. I understand all the NIST raw data could be made available if a researcher wanted to do something with it. Unfortunately, I'm fully engaged starting netBlazr Inc., so I have find other researchers' works or wait for someone to look into this.

The one thing that's seems clear from the data I have seen is that masonry does not significantly absorb RF, i.e. convert it to heat. It mostly scatters RF.

Henning

Good post about the value of MIMO and how it can make digital radio usage much more efficient, especially for the high frequencies. Just as a comment, both beamforming and beamstearing have been used for military radar for quite some time, https://secure.wikimedia.org/wikipedia/de/wiki/Mammut_%28Radar%29 . Its just that its getting cheap enough for consumer hardware now.

But I am not sure if its possible to get "secondary/whitespace" usage before the primary usage is defined. For getting a "whitespace licence" you have to show that you do not interfere with the primary user, which sounds impossible with a frequency not allocated.

brough

Thanks Henning. Indeed phased array concepts go way back! Nobel Laureate Karl Ferdinand Braun demonstrated the concept in 1905 and I understood AM broadcasters were using phased arrays in the 1930s (although I can't find the reference right now). But you are correct that it was radar and WWII that drove electronic beam steering. What's exciting is that the MIMO specs, e.g. 802.11n, 802.11ac and similarly in WiMAX & LTE, provide for transmit beamforming/beamsteering while the MIMO calculations inherently provide the equivalent of receive beamforming/beamsteering. The 802.11n & 802.11ac specs combined with Moore's law pretty much insure we will get consumer priced beamforming/beamsteering within a few years.

"Secondary" access is already well established. Many of the amateur radio bands have worked on that basis for decades. In the US, the UNII-2 (5.25-5.35 GHz) and UNII-extended (5.47-5.725 GHz) bands are secondary use, based on "DFS" which is a radar sensing scheme. TV white spaces introduce database lookup as another way to determine when a secondary user can access a band without interfering with the primary licensee. There's also the 3650-3700 MHz band in the US. While the 802.11y specification was originally developed for Wi-Fi in the 3650 MHz band, the 11y committee produced a very general scheme that allows for sensing, data base look up and for master & slave devices so the extra cost of sensing and/or database look up can be separated, thus reducing the average cost of a system of devices.

Carlson Wireless

As the spectrum database management model matures, perhaps it will become easier to convince those license holders that it's safe to share. Thanks for this article, Brough.

Steve Crowley

Hi Brough. I think you're on the right track here. I agree we're in the early days of MIMO and adaptive technologies. The article implies that MIMO was the first to constructively use multipath interference, but OFDM and adaptive equalization do that also. MIMO adds significant additional gain.

Regarding the loss mechanism with water vapor, in reference 2 I see the peak of attenuation for water just above 20 GHz. Is that due to resonance? Elsewhere, I assume attenuation from water vapor would be by the same mechanism as from other materials, dielectric heating caused by molecule dipole rotation.

The FCC has a proceeding underway on dynamic spectrum access. Maybe there will be some more flexibility from that. We do need more flexible use of spectrum.

Many higher frequencies are in government bands, many of which I suspect are underutilized. NTIA is looking at repurposing some of those. Unfortunately, a GAO report earlier this year concluded, basically, that NTIA's databases are a mess, and no one really knows what's going on in the government bands. We need a good government database, informed by an inventory of spectrum assignments and usage, in my opinion.

Given a lack of appropriate regulation, if one wanted to do more now, one could try for a waiver of the rules, or try to operate under experimental authority until permanent rules are in place, not that either of those are easy.

brough

Thanks for the comment Steve. I agree OFDM with cyclic extension and channel adaption makes constructive use of the multi-path energy, as does a CDMA system with a rake receiver. But somehow, these systems strike me as doing the best possible mitigation whereas MIMO does all this while also leveraging multi-path to increase capacity. But yes, you are correct.

Yes, the first water peak is at 22 GHz. Water vapor has a complex spectrum with over 64,000 spectral lines listed in the 2007 HITRAN database. They are all the result of the incoming electromagnetic energy exciting different vibrational modes of the molecular bonds. It's H2O, so there are two bonds, each of which can stretch, rotate, bend, etc., either symmetrically or anti-symmetrically with the other bond.

There's a good graph of atmospheric absorption by water vapor and oxygen here:
http://www.rfcafe.com/references/electrical/atm-absorption.htm

One possible way to flesh out who claims what spectrum might be to declare it all will be made available three years from today for secondary use subject to database lookup and power limits. Then give federal agencies and others two years to object and up to three years to submit receiver location information to the database contractors. At a minimum, such a proposal might help NTIA get their databases in order. :)

The comments to this entry are closed.

My Photo

Search this Blog

Subscribe by Email

March 2014

Sun Mon Tue Wed Thu Fri Sat
            1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

Technorati


Site Meter

Upcoming Travel & Conferences


Twitter Feed

Become a Fan