The independent blog on spectrum policy issues
that welcomes your input on the key policy issues of the day.

Our focus is the relationship between spectrum policy
and technical innnovation.

cognitive radio

MS Spectrum Observatory: Seeing Things NTIA Doesn't Want to See


In August 2010 we published a post here entitled “Does NTIA Have the Bureaucratic Courage to Measure Occupancy of 225-400 MHz?” At the time we used the following chart to show bands NTIA had measured and publicly reported on and the middle band they had carefully ignored.
The 225-400 MHz band is home to many federal systems, but is best known for military aircraft communications. In other countries it is used for civilian aircraft communications and in countries where the US has bases, e.g. Japan and NATO countries, there is tenuous sharing. A fact of US geography is that with the notable exception of San Diego there are no large military airfields in major cities. Thus in geography military spectrum use in this band is generally orthogonal to civil needs for spectrum - that is you rarely see a heavy demand for both military aircraft spectrum and civil spectrum users in the same place.

Long time readers may recall that your blogger has been advocatinginterruptible spectrum” use here for a long time. Interruptible spectrum is similar to interruptible electricity. One has access to the resource, but can be preempted from time to time when there is great demand for the resource from users with higher priority.

If your view of cellular communications is people calling 911 then this doesn’t make much sense. But all voice communications is a shrinking faction of cellular traffic and 911 traffic must be microscopic. The great growth in demand for mobile broadband is basically for discretionary services, not safety related services. While the cellular community would like to guarantee all users a high quality of service, there frankly isn’t the spectrum available to do so. Other utilities offer pricing schemes to affect demand with approaches like peak pricing in addition to offering interruptible service at a discounted price. The cellular industry just hasn’t been thinking this way.

But similarly NTIA has not been thinking “out of the box” about efficient use of the natural resource at 225-400 MHz despite new White House thinking. The June 14, 2013 Presidential Memorandum -- Expanding America's Leadership in Wireless Innovation stated:

“Although existing efforts will almost double the amount of spectrum available for wireless broadband, we must make available even more spectrum and create new avenues for wireless innovation. One means of doing so is by allowing and encouraging shared access to spectrum that is currently allocated exclusively for Federal use. Where technically and economically feasible, sharing can and should be used to enhance efficiency among all users and expedite commercial access to additional spectrum bands, subject to adequate interference protection for Federal users, especially users with national security, law enforcement, and safety-of-life responsibilities.”

Does 225-400 MHz meet this test of economic and technical feasibility? The lack of official data on its occupancy discourages commercial interest. (Of course the cellular industry’s impossible goal of only getting spectrum that is available 24/7 in all US territory and can use Chinese-made electronics also discourages interest .) There has been some data available in public reports from Shared Spectrum for several years. (See Note 1 below)The following plot was in an early post here.
In case the August 2004 date doesn’t ring a bell, this data was taken during the 2004 Republican Convention in New York City 3 years after 9/11 when an unprecedented federal law enforcement and military presence was there to protect national leaders. The third plot shows that spectrum use in 225-400 MHz was negligible despite this high federal presence. Other Shared Spectrum publicly available measurements are at their website and show a similar pattern.

There is also data available from the IIT Spectrum Observatory in Chicago which has been making observations since 2007 and storing them. Unfortunately there is not a simple way for researchers to access this data without making detailed arrangements with IIT.

We recently became aware of another resource, the Microsoft Spectrum Observatory which makes observation between 30 and 6,000 MHz in Brussels, Redmond, Seattle, and Washington DC. Unlike the IIT measurements, these are available for online access to anyone!

So how busy is 225-400 MHz in Washington DC and Seattle? In the 2 charts below I picked for the whole day of 1/21/13 in DC, the day of President Obama’s second inauguration and an exceptionally high federal presence. In the case of Seattle I picked 6/4/13, a day I happened to have been there for a conference which I recall was a bright warm sunny day unlike what one expects in Seattle.
Spectrum Occupancy Washington DC 1/21/13 24 hour data (225-400 MHz)

Spectrum Occupancy Seattle WA 6/4/13 1100-1200 (225-400 MHz)

So the cellular industry is not interested in this band because it wants 24/7 nationwide access and the ability to use Chinese-made electronics that meet global standards. Why isn’t FCC and NTIA interested? Repeatedly the problem I hear is that it would complicate the job for DoD spectrum managers stationed overseas who have to explain to our allies why dedicated military channels are needed in this band. The fear is that if DoD doesn’t need 24/7 access to this band in the US, why do they need access in Japan and Europe? This is a real issue but I think it is time to stop holding US spectrum use hostage in order to simplify the job of DoD spectrum managers overseas. The issue here is not reallocating this spectrum to 24/7 commercial use and not using it for 5GHz-like “dynamic spectrum access” that depends on sensing, but rather have a dynamic database system that updates cellular base station in real time which frequencies can be used on a fail-safe basis. Normally base station would have access to much of the band, but as military traffic start building up in an area, blocks would be cleared in anticipation of more traffic so the military traffic would always have what it need plus an additional block of spectrum for rapid surges.

Note 1: There is no detailed generally accepted way to measure occupancy so it is impossible to make precise comparison of measurements. Many decisions have to be made before data is collected and those decisions affect the numbers. Some possible decisions increase apparent occupancy, some decision decrease it. I vividly recall climbing Mt. Fuji in Japan about 15 years ago with my trusty 2m amateur radio. At the top of 12,388' (3,776 m) Mt. Fuji all amateur radio frequencies in the 2m band are fully occupied since the radio horizon includes most of the main island of Honshu! Similarly sensitivity/noise figure, filtering, and 3rd order intercept point decisions affect apparent occupancy. But that said, measurements with very low occupancy show that there is a real issue that should be investigated to confirm or deny whether spectrum is available. Note also that spectrum occupancy measurements usually miss satellite uplink and some downlink transmissions due to their nature. Measurements of full duplex bands with base stations also tend to see the base station downlinks and due to siting can miss the corresponding uplink traffic.

Note 2: How might 225-400 be shared with the private sector? It is clear that any use mobile-based spectrum sharing decision system is not secure enough for the national security uses of this band. Also no listen-before-talk (LBT) cognitive radio system would be reliable enough either. But a base station-based system, typical in today’s cellular systems would place the spectrum selection system in the hands of a few carriers. The carrier spectrum election system could then in turn be controlled in a fail/safe way by DoD spectrum managers to make sure that at a given moment the spectrum is adequate to handle at least (100 + x)% of the authorized DoD traffic. As the DoD traffic approaches the amount set aside for them at a given time in a given area the amount set aside should be increased to maintain a positive margin of available spectrum above demand. Some may argue that the physics of this sharing is similar to the 1755-1850 MHz sharing considered by NTIA’s CSMAC. The D/U ratios discussed there were computed using a primitive propagation that IRAC prefers because it is very conservative. A more balanced consideration of sharing issues under the terms of the 6/13 Presidential Memo would likely give differing answers.


See http://www.marcus-spectrum.com/Blog/files/SpecMeasCorr414.html


DARPA Starts to Tackle Radar/Comm Sharing

Radar/communications sharing has been a recurring topic herein. The growing demand for “beachfront” mobile spectrum to assure economic growth and provide for key wireless public safety systems require a reexamination of the large exclusive allocation for “radiolocation” - radar in spectrum policy jargon.

Radiolocation alloc
Traditionally federal radars were given large exclusive bands as primary allocations and the federal owner of the radar, usually FAA or a military service, was able to make design tradeoffs between system cost/complexity and the amount of spectrum utilization they achieved. Since spectrum was a free cost to the designer there were few incentives to minimize spectrum footprint other than what the owner’s IRAC representative was able to get them in the “smoke filled room” that is IRAC and its subcommittees. While NTIA in theory had oversight over this process, it most cases it did not take an active role if the IRAC members were able to agree among themselves and there was no strong private sector pressure - usually the cased since IRAC deliberations and even agendas were secret.

When sharing was proposed, the same IRAC was able to dictate sharing conditions to FCC with the functioning of the radar as originally designed being the paramount issue. Overprotection that limited private sector sharing was of little concern to IRAC and the technical criteria development was often shrouded in secrecy to protect the technical details of the radars involved. This is shown in the case of the 5 GHz U-NII sharing where we have the bizarre sharing conditions of § 15. 407:

§ 15.407(h)(2) Radar Detection Function of Dynamic Frequency Selection (DFS)
A channel that has been flagged as containing a radar system, either by a channel availability check or in-service monitoring, is subject to a non-occupancy period of at least 30 minutes. The non-occupancy period starts at the time when the radar system is detected..Radar Detection Function of Dynamic Frequency Selection (DFS). U-NII devices operating in the 5.25-5.35 GHz and 5.47-5.725 GHz bands shall employ a DFS radar detection mechanism to detect the presence of radar systems and to avoid co-channel operation with radar systems. The minimum DFS detection threshold for devices with a maximum e.i.r.p. of 200 mW to 1 W is −64 dBm. For devices that operate with less than 200 mW e.i.r.p. the minimum detection threshold is −62 dBm. The detection threshold is the received power averaged over 1 microsecond referenced to a 0 dBi antenna. The DFS process shall be required to provide a uniform spreading of the loading over all the available channels…

Non-occupancy Period.
A channel that has been flagged as containing a radar system, either by a channel availability check or in-service monitoring, is subject to a non-occupancy period of at least 30 minutes. The non-occupancy period starts at the time when the radar system is detected.

According to these 2 sections if an unlicensed device detects any signal on the channel greater than - 64 dBm (non techies: that is less than a millionth of a milliwatt!) it must avoid using that channel for the next 30 minutes! Double checking to see if it is a false alarm due to noise is not allowed by this rule.

(It is ironic that the recurring interference to NOAA’s TWDR weather radars is not addressed here since TDWR is a doppler system unlike FAA radars. It appears that DoD and FAA dominated the secret IRAC deliberations on the criteria that was dictated to FCC by NTIA and did not pay much attention to NOAA’s radar technology.)

To NTIA’s credit they did choose radar/communications sharing as the theme of the 2011 International Symposium on Advanced Radio Technologies, their annual public technical conference in Boulder CO. The conference was entitled “Developing Forward Thinking Rules and Processes to Fully Exploit Spectrum Resources: An Evaluation of Radar Spectrum Use and Management”. In the proceedings, is a paper (p. 214) from your blogger entitled “Thoughts on Radar/Communications Spectrum Sharing”. This paper breaks from the static precedent of U-NII radar sharing and suggested time dynamic sharing, noting that while 1000 ms/s spectrum access was essential when circuit switched voice dominated mobile traffic, today’s packet switched data is the dominant traffic and packet switching technology can readily make reliable networks our of time intermittent links.

The paper was discussed at the conference and here are two comments from the proceedings:

A second idea that is very important, first advocated for by Michael Marcus, is about unlicensed low power uses and it has spawned a revolution of technological developments that take advantage of that possibility. -- Phil Weiser, U. Colorado (Formerly senior advisor for technology and innovation to the National Economic Council at the White House)

I also wanted to take just a minute to recognize that Mike Marcus, who was originally going to be on this panel but couldn’t make it in the end, also put together a paper that talks about some sophisticated techniques for sharing with radars. -- Julius Knapp, FCC

DARPA logo
Following up on the conference and perhaps on the ideas of my paper is the new DARPA program, Shared Spectrum Access for Radar and Communications (SSPARC) in support of Broad Agency Announcement (BAA) DARPA-BAA-13-24 which will have a Proposer's Day on February 26th at DARPA - registration required:

The purpose of this conference is to provide information on the SSPARC program; promote additional discussion on this topic; address questions from potential proposers; provide a forum for potential proposers to present their capabilities for teaming opportunities; and following the conference, conduct pre-scheduled one-on- one meetings with the Program Manager as described below.

PROGRAM OBJECTIVE AND DESCRIPTION: DARPA anticipates releasing the BAA for SSPARC with the objective described in this section. The objective of the SSPARC program is to improve radar and communications capabilities through spectrum sharing. SSPARC develops technology applicable to spectrum sharing between military radars and military communications systems, and between military radars and commercial communications systems. SSPARC includes work on spectrum sharing systems and separation mechanisms, supporting technologies that improve performance when sharing spectrum, theoretical performance limits and grounded design techniques, and relevant regulatory topics.

It is gratifying to see that this concept is moving along towards more detailed study after being ignored for a long time.

Now if only the cellular industry would address the topic of using intermittent access to radar spectrum fro CMRS with an open mind rather than insisting on an unrealistic goal of meeting all their spectrum demands solely by reallocation.


The DARPA BAA is now posted here. From the introduction:

The Shared Spectrum Access for Radar and Communications (SSPARC) program seeks to improve radar and communications capabilities through spectrum sharing.

Spectrum congestion is a growing problem. It increasingly limits operational capabilities due to the increasing deployment and bandwidth of wireless communications, the use of net-centric and unmanned systems, and the need for increased flexibility in radar and communications spectrum to improve performance and to overcome sophisticated countermeasures. Radar and communications jointly consume most of the highly desirable spectrum below 6 GHz. SSPARC seeks to develop sharing technology that enables sufficient spectrum access within this desirable range for radar and communications systems to accomplish their evolving missions.

The SSPARC program seeks to support two types of spectrum sharing.

  • Spectrum sharing between military radars and military communications systems (“military/military sharing”) increases both capabilities simultaneously when operating in congested and contested spectral environments.
    • Spectrum sharing between military radars and commercial communications systems (“military/commercial sharing”) preserves radar capability while meeting national and international needs for increased commercial communications spectrum, without incurring the high cost of relocating radars to new frequency bands.

UPDATE 2/26/13

Today’s public meeting on the DARPA BAA was a great success with about 150 people present! There clearly is great interest in the topic in the R&D community.

Is there any interest in using such spectrum from the cellular establishment or will they continue their negative reaction to the PCAST report and demand 500 MHz of new spectrum subject to
  • nationwide availability
  • 24/7 1000 ms/s availability
  • compliance with 3GPP band plans?

But as the PCAST report points out this is an unrealistic goal in a country with the largest most information intense military in the world. The cellular industry aimed too high in the “booster amplifier” rulemaking, Docket 10-4. Should they be more reasonable here?

UPDATE 3/15/13
SSPARC_144_144The SSPARC program now has the logo shown at left. It carefully does not show any physical system in order to be open minded about possible designed. DARPA has now posted on their website presentation from the Proposer’s Day. Of particular interest is a presentation from DoD spectrum managers showing how much spectrum is used by radar today.

IEEE Dynamic Spectrum Access Workshop in DC 9/19/11

The IEEE Communications Society will sponsor a workshop on dynamic spectrum access (DSA) technology and related policy issues in Washington on 9/19/11. IEEE states:

“This workshop brings together thought leaders from the government, regulatory bodies, Department of Defense, industry, standards bodies, and academia to discuss and identify inter-disciplinary opportunities and challenges in dynamic spectrum access wireless networking. This unique workshop program contains only interactive panel sessions in order to foster collaborative discussions among the participants. Break-out sessions also provide a forum for one-to-one interaction.”

The program is now available and registration is open.

Radar/Communications Spectrum Sharing: ISART 2011


On July 27-29 NTIA will host the 12th Annual International Symposium on Advanced Radio Technologies (ISART) at its Boulder, CO Institute for Telecommunications Sciences. The theme this year is “Developing Forward-Thinking Rules and Processes to Fully Exploit Spectrum Resources” with a special focus on radar bands.

For the third time since his retirement from FCC, NTIA was kind enough to invite your blogger to speak at this important meeting, but due to a conflicting family event, he is unable to attend. However, in view of the importance of this issue, he volunteered to produce a written paper on the topic to help stimulate discussion. Here is a link to that paper.

The paper starts by stating the need for new spectrum to speed economic growth which is important for both our society and for national security. Spectrum allocation should not be viewed as a zero sum game, but it is critical to develop innovative sharing techniques to get the maximum use of this limited resource. Since radar systems are a large user of spectrum and are difficult to share with using conventional approaches, this is a very timely conference.


The paper then urges FCC and NTIA to be more transparent about ongoing 5 GHz U-NII interference problems with the FAA’s TDWR weather radar. We previously mentioned the mysterious San Juan TDWR case involving interference from an AT&T-operated Motorola unit and which is puzzling based on the public record to date. NTIA has prepared a report on the San Juan case, but it mainly confirms that the DFS function in the AT&T/Motorola unit wasn’t working without explaining why. Another NTIA presentation from last year’s conference confirms that while some cases are caused by nonfunctional DFS, some have been caused by compliant DFS systems used in WISP systems, necessitating an unusual “voluntary” FCC agreement with WISPA to effectively coordinate such systems to prevent such interference. The paper urges FCC and NTIA to explain more what the root causes of these problems are so that future systems can be designed to avoid them.

The main part of the paper advocates joint design of new radar systems with communications experts in order to maximize spectrum sharing and consider financial cost sharing of features that facilitate sharing subject to the radar mission needs. Just as the stealth bomber design involved a unique team of aeronautical engineers and EEs who could trade off flying issues with radar visibility issues, joint design of radar/comm systems may well result in sharing breakthroughs. While current legislation does not allow this type of cost sharing, it is not beyond the reach of new legislation that has been discussed. The paper points out that while full duplex paired spectrum with “24/7 and 1000 ms/ 1 s” time availability has been the norm for commercial systems, the decline of voice minutes and the domination of packetized traffic means that partial time availability, synched with radar rotation, could result in productive access to radar spectrum. While nonmilitary backlobe radio have not improved in 40+ years, advances in radio astronomy antennas indicate that new designs can significantly decrease backlobe levels and facilitate sharing. Such designs are expensive, but cost sharing could address that.

if you are interested, here, again, is the link.


Cognitive Radio Conference in Singapore

Greetings from Singapore, location of IEEE DySPAN 2010 - the world’s premier cognitive radio conference.

Despite the poor economy, there are 253 attendees registered as the conference starts, including 47 from US. No one from FCC or NTIA came, although FCC alums include Bill Luther, Paul Kolodzy, Doug Sicker (the conference cochair), and myself.

During the keynote session, Dr. Geok Leng Tan, CTO of the Information Development Authority of Singapore, the local “FCC”, proudly announced a new program to encourage field testing of TV band white space devices in Singapore. This is a bold approach for a country only a few km from neighboring Malaysia.

In its announcement IDA stated

1.1 In Singapore, the spectrum of radio frequencies currently allocated for use, on a primary basis, for the provision of terrestrial broadcast services in the Very High Frequency (“VHF”), between 174 to 230 MHz, and the Ultra High Frequency (“UHF”) bands, between 494 to 790 MHz, are not all in use at any given time. The spectrum not in use for the provision of broadcast services in Singapore consists of both interleaved and contiguous spectrum blocks. The unused spectrum is known as „white spaces‟. These white spaces are situated below 1 GHz, and the propagation characteristics at these frequencies allow radio signals to travel long distances and penetrate buildings easily.

1.2 IDA welcomes the conducting of trials for white space technology in Singapore, by any interested party, to explore the various spectrum environments and regimes that white space technology could operate in. To facilitate the investigation and testing of white space devices and systems in Singapore, and to ensure continued protection of licensed services in Singapore from possible interferences that may occur during the trial period, IDA has developed a set of terms and conditions and the application procedures for interested parties who wish to conduct white space trials.