August 25, 2000

Steve Jones
Office of Spectrum Management
National Telecommunications and Information Administration
Room 6725 HCHB
1401 Constitution Avenue, NW
Washington, D.C. 20230

RE: Request for Comments on GPS/UWB Measurement Plan

Dear Mr. Jones:

Thank you for the opportunity to provide comments on NTIA's GPS/UWB Measurement Plan. This is an important effort in view of the increasing worldwide reliance on GPS signals for security, safety, scientific research, and economic growth. UWB technology has been the subject of government research for many years and now appears to have promising commercial potential, albeit with novel technological and regulatory issues.

With regard to the questions for public comments:

Question 1: Are the candidate GPS receivers identified in the measurement plan representative of the different technologies and user applications?

• The candidate list is a good starting point, but given the multitude of GPS applications, it is not complete. In particular, receivers used for E-911, autonomous vehicle positioning and three-dimensional machine control should be included for their challenging terrestrial requirements. Attachment 1 provides a representative list of GPS applications submitted to working groups at the International Telecommunications Union.
• Given the importance of E-911 and evolving decisions on 3G wireless infrastructure, the testing of GPS embedded in wireless phones be a of particular interest to both NTIA and the FCC.
• A related issue is whether NTIA will have the time and resources to test even those receivers listed. If not, the measurement plan cannot be considered complete by its own standards.

Question 2: Are the UWB transmission system parameters identified in the measurement plan representative of the parameters for UWB transmission systems envisioned for use by the public?

• No. Only UWB signals with pulse widths of 0.5 nsec duration are generated. Technical literature suggests UWB signals can have both positive and negative harmonic components and can vary from 0.5-10 nsec, and have varying amounts of damping depending on the driving signal and specific transmission antenna.
• It is unclear what is meant by "envisioned for use by the public." One interpretation is that only UWB devices with characteristics specifically tested by this measurement plan could be considered for public use (licensed or not).

Question 5: Is a performance metric of time to reacquire a satellite applicable to GPS receivers used for terrestrial applications (e.g., public safety)? If so, what are the associated performance criteria?

• Yes. There may be a wide variety of associated performance criteria based on specific applications and operational scenarios. In general, a conservative assumption should be that harmful interference is that in which intentional emissions leads to the loss of one or more signals from the GPS satellite constellation. Then the time to reacquire would indicate whether the loss could be tolerable - that is, lead to no degradation of the GPS application's functionality.

Question 7: What are the performance metrics and associated criteria for GPS receivers used for surveying, maritime, and recreational applications?

• Performance metrics are most effectively determined by users and their respective organizations, in conjunction with GPS manufacturers, applications developers, and systems integrators meeting user needs. Thus those groups should determine criteria for surveying, maritime, and recreational applications in the absence of government or international agreement.

• In Question 6, it was stated that "A reacquisition time of 1 second has been proposed by at least one GPS receiver manufacturer for terrestrial applications." Reacquisition time is often related to integrity requirements for applications involving public safety. One second "time to alarm" requirements for intelligent transportation systems (including automobile, bus, and rail uses) have been reported by the NTIA for many years. The criticality of reacquisition time can vary depending on the availability and practicality of additional information sources such as inertial guidance.

With regard to the draft measurement plan:

Page 1 -

UWB is a recent and novel technology for commercialization. It has been used for specific military applications, especially radar, for many years but subject to strict controls. It should be noted that the intentional transmission of uncontrolled and unfiltered signals across extremely wide bands was done in the early years of the 20^th century by radio pioneers. However, the interference problems created by such activities led to the formation of the FCC's predecessors and adoption of more disciplined, technical and regulatory approaches such as the use of assigned frequencies and channels.

It is unproven opinion that UWB can perform useful telecommunication functions that make them "very appealing for both commercial and government applications." Rather that view is one that will hopefully be determined by public testing and demonstration elsewhere. Experiences in military applications are not automatically analogous to commercial applications.

The statement that the average power levels of UWB devices are "low enough to be authorized under the unlicensed device regulations" is similarly an unproven assertion that may or may not be borne out by rule making. Other factors that must be considered include emissions into restricted bands and peak power limits. Some countries such as Japan, do not even have "unlicensed" device regulations, but instead have rules for licensing "low power" devices. Thus the outcome of FCC rule making in the area of UWB may result in differing international treatment that is more or less lenient toward UWB devices.

Part 15 regulations were formulated to address narrowband, continuous wave systems while UWB devices are wideband, discontinuous wave systems. Although an average power measure correlates with the interference potential of continuous wave systems, peak power is a more indicative characteristic. A transient signal with rapid change in field strength may have a broad and flat power density spectrum, but be capable of electronic upset or even damage. To give a simple example:

• 1 joule of energy delivered continuously and evenly over one second is 1 watt of power
• 1 joule of energy delivered discontinuously over 1 nanosecond is 1 gigawatt of power
• 1 joule of energy delivered discontinuously over 1 nanosecond and measured at a rate of 1 sample per second will show up at 1 watt of power.

Fast risetime pulses can produce multiple harmonic response in a narrowband receiver and have thus been considered for their potential as intentional jammers in weapon systems.

Page 2:

Factual corrections: "GPS-based public safety systems and services are being fielded" and "…fully dependent upon GPS signal reception worldwide." It should also be mentioned that U.S., and allied militaries are reliant on GPS and becoming increasingly so such that any degradation of GPS signals would adversely affect national and international security.

"…timely policy and regulatory decisions must be made are being considered." It has not been shown in policy or law that regulatory decisions "must" be made. Rather policy and law has appropriately placed an emphasis on protecting GPS signals from interference for reasons of national security and public safety.

"…projected to proliferate rapidly by advocates of UWB applications." NTIA has not done such a projection, rather such projections have come from firms hoping to benefit from favorable regulatory treatment.

Page 2 -

Operational scenarios in which UWB and GPS equipment is in proximity are not the only operational scenarios of interest. The NTIA test plan should consider instances where UWB devices operate in dense communication networks over wide areas. If there were several such overlapping networks, as some advocates envision, how might the aggregate noise environment experience by GPS change?

Page 3 -

Others stressing operational scenarios should include:

• precision machine control, as in autonomous navigation of large machinery where rapid updates of precise vehicle attitude as well as position are needed;
• use of GPS to enable E-911 and in conjunction with time-difference of arrival techniques;
• precise time transfer which is key to so many telecommunications systems; and
• positive train control with accuracy requirements of 2 cm and integrity requirements similar to aviation as expressed by the Federal Railroad Administration.

Link budgets need to consider not just GPS and UWB, but other noise sources allowed to operate in the band such as existing unlicensed devices, mobile earth stations, and other RNSS systems such as GPS augmentations and the Galileo system being developed by Europe.

The term "any necessary restrictions" may involve both technical and operational components and what assurances are possible to ensure the restrictions are complied with. This includes the prevention of unintentional or accidental misuse, for example, when damage to the UWB transmitting antenna results in significant changes to the emitted signals.

Objectives to identify "UWB signal parameters to be considered," develop a "GPS/UWB measurement methodology," and develop measurement procedures to assess the potential for signal source and aggregate UWB interference are certainly valuable and needed. However, only one kind of UWB source and three "aggregated" sources are discussed. Thus the aims of the test plan are not fully supported by the time and resource limitations now imposed on what the test plan can actually accomplish.

Page 4 -

The intended use of "high speed time domain equipment" is an excellent step, but there is no description of such equipment or why its use is necessary (i.e., why spectrum analyzers alone are inadequate). It is also important to measure the "spectral content and characterization" of pulse amplitude distribution over time" but again, no indication of the availability of test equipment to do this.

Page 7 -

Actual GPS receivers should also be compared to minimal ITU standards for GPS (space-to-Earth) receivers. Similar standards, to a high degree of detail, need to be developed in the future for UWB devices to facilitate interference studies. Precise test programs and analysis are both needed for a full evaluation of GPS and UWB interference issues. In a full-up evaluation, one performs a theoretical analysis, i.e., develops a performance model. The measurements are intended to validate the model. Then, as a final step, the model is exercised under a very broad set of conditions, e.g., corresponding to the geometries of interest between the various signal and interference sources and receivers, the effects of other ambient interference sources, the effects of the particular antenna installation and configuration, and the range of propagation effects. The reasons the work should be done this way are that:

Page 9 -

Only UWB signals with pulse widths of 0.5 nsec duration are generated. Technical literature suggests UWB signals can have both positive and negative harmonic components and can vary from 0.5-10 nsec, and have varying amounts of damping depending on the driving signal and specific transmission antenna.

Pulsewidth and shape may have an effect on measurement if UWB signals exist above or below the GPS band and overlap into the band. If GPS can be protected by "notch filtering" for example, a more precise characterization of UWB parameters will be needed.

UWB power level per se may not be the only variable the leads to effects on GPS receivers. How does NTIA propose to explore the space of possible parameters and characterize them?

The test plan refers to an average power density as measures in 20 MHz bandwidth (dBm/20 MHz). On average, the sampling rate of present spectrum analyzers is 20 MHz, with state-of-the-art at about 30 MHz. Such a sampling rate is inadequate to provide an accurate measure of signal amplitude and thus measure power. A fast time-domain sample-and-hold oscilloscope should be used to determine signal temporal duration and phase and thus augment the spectrum analyzer to get a reliable power measurement.

Page 10 -

GPS receivers process each channel independently but all satellites in view are valuable. In practice more than four are needed since those with the best geometry are continuously and actively chosen. Worst case analyses and conservative assumptions are warranted in analyzing the potential for interference just as the United States did in studies conducted for Working Party 8D in the most recent ITU study cycle.

Only limited samples of GPS receivers are to be measured. Thus is important to explain more fully why 2dB was chosen as the appropriate margin for unit-to-unit variability.

Page 13 -

In general, GPS performance criteria have not been defined the by the U.S. government. They do exist for aviation within ICAO SARPs (Standards and Recommended Practices) and from GPS manufacturers in meeting the needs of their diverse customers (including the U.S. Government). U.S. policy, law, and international agreements have all stressed the need to protect GPS signals from harmful interference for reasons of national security and public safety. The U.S. government does not decide that some users are worthier of protection than others - that is, it does not designate victim applications.

Page 15 -

How might multiple UWB emitters vary among themselves to produce aggregate interference effects? In some cases, one might image coordinated emissions that add like noise. In others, uncoordinated UWB networks could create noise that is greater than any single emitter could.

Page 21 -

UWB peak power observed or filters used should be recorded. How might spectral lines be observed and recorded? How stable and repeatable are UWB devices tested likely to be?

The caveats in this section include the use of an "extremely aggressive schedule" that "severely limits the scope of these tests" are well taken and justified. The current schedule does not appear to include time to do analyses, write a report, and secure peer review or wider public comment. NTIA would be well advised to secure an extension with the FCC in order to write a thorough report and not take any action (including use of "provisional" rules) until there has been complete testing and a full and open technical debate.

As stated as the beginning, the NTIA test plan represents an important effort with profound technical and regulatory implications. The necessary time, care, and resources should be provided to protect the public that now relies on GPS while finding means for beneficial UWB applications to be safely encouraged.

Respectfully submitted,

Scott Pace
Senior Policy Analyst
RAND Science and Technology Policy Institute
1200 South Hayes Street
Arlington, VA 22202-5050
(703) 413-1100

Attachment