BEFORE THE
FEDERAL COMMUNICATIONS COMMISSION
Washington, D.C. 20554

In the Matter of
Wireless E911 Location Accuracy
Requirements

PS Docket No. 07-114

T-MOBILE USA, INC. COMMENTS ON
THIRD FNPRM ON LOCATION ACCURACY

Steve B. Sharkey
Eric Hagerson
T-MOBILE USA, INC.
601 Pennsylvania Ave., NW
Washington, DC 20004
(202) 654-5900

John T. Nakahata
Kristine Laudadio Devine
HARRIS, WILTSHIRE & GRANNIS LLP
1200 Eighteenth Street, NW
Suite 1200
Washington, DC 20036
(202) 730-1300
Counsel to T-Mobile USA, Inc.

May 12, 2014

 TABLE OF CONTENTS
I. 

INTRODUCTION AND SUMMARY ............................................................................................ 1 

II. 

LOCATION ACCURACY IS BEST IMPROVED THROUGH SUPPORT OF INDUSTRY EFFORTS
AND MULTI-STAKEHOLDER TESTING. ................................................................................... 6 

III. 

A. 

Current and Near-Term Technologies Will Continue to Improve Location
Estimates for Wireless 911 Calls. ............................................................................ 6 

B. 

The Commission Should Move to Establish a Permanent Indoor Test Bed and
Rely on CSRIC’s Recommendations Regarding New and Emerging Location
Technologies ............................................................................................................ 8 

TECHNICAL AND ECONOMIC FEASIBILITY AS WELL AS PUBLIC SAFETY’S NEED FOR
“TRULY ACTIONABLE” LOCATION INFORMATION MUST GUIDE THE COMMISSION’S
DECISIONMAKING. ................................................................................................................ 9 
A. 

The Indoor Location Proposals Ignore the Results of the CSRIC Test Bed. ........ 11 
1.  NextNav’s Beacon Technology Is No Panacea. ............................................. 12 
2.  Other Location Technologies Are Similarly Challenged. .............................. 17 

B. 

The Indoor Location Proposals Establish Unrealistic Timeframes Given the
Need to Test and Deploy Equipment as well as for Handset Turnover. ............... 18 

IV. 

MANY OF THE COMMISSION’S PROPOSALS APPEAR TO CONFLICT WITH ONE ANOTHER
AND CREATE INCONSISTENT OR CONFUSING OBLIGATIONS................................................ 20 

V. 

THE COMMISSION SHOULD REFRAIN FROM MODIFYING THE EXISTING LOCATION
ACCURACY FRAMEWORK FOR OUTDOOR CALLS. ............................................................... 21 

VI. 

CONCLUSION ...................................................................................................................... 23 

 I.

INTRODUCTION AND SUMMARY
T-Mobile USA, Inc.1 (“T-Mobile”) believes that continued work regarding Enhanced 911

(“E911”) capabilities is critical to ensuring that wireless customers continue to have access to the
best possible location technology in the event of an emergency. And T-Mobile stands second to
none in its commitment to the safety of our customers. Just as we have aggressively thought
outside the box to remove consumer pain points for wireless service, we have also worked with
and are ready to continue to work with the Commission, other carriers, public safety, and other
important stakeholders—including local governments, premises owners, technology developers,
and manufacturers—to think creatively about the challenges associated with generating location
estimates indoors. A creative and inclusive approach is important for all improvements to
location accuracy, but it is particularly critical with respect to the Commission’s stated goal of
ensuring that public safety is provided with a truly “actionable location” for wireless 911 calls
placed from inside large multiunit and multistory buildings.2 That type of creative solution is not
yet on the table.
The plain fact is that, while wireless carriers can continue to make incremental
improvements, they cannot by themselves deliver fully “actionable” horizontal and vertical
indoor location estimates; when limited to technologies that wireless carriers can incorporate in
their networks or handsets, a wireless carrier can only generate estimated center coordinates for a
two- or, in some instances, three-dimensional probabilistic search radius. Put more simply,
regardless of whether the proposed new accuracy standards are adopted, if the Commission and
public safety rely solely on wireless carriers to provide location information, it will always be
1

T-Mobile USA, Inc. is a wholly-owned subsidiary of T-Mobile US, Inc., a publicly traded
company.

2

Wireless E911 Location Accuracy Requirements, Third Further Notice of Proposed
Rulemaking, FCC 14-13, 29 FCC Rcd. 2374 (2014) (“Third FNPRM”).

 possible that the caller is on some other floor or in a different building. To attain truly actionable
indoor locations requires buy-in and development from all stakeholders—not just wireless
carriers, but also public safety, handset manufacturers, location technology vendors, mobile
operating system providers, state and local governments who regulate building codes, and,
perhaps most critically, premises owners. “Smart buildings” should be more than just wired—
they should also be capable of providing actionable locations to any mobile device (including
Wi-Fi-only devices).
In the nearer term, T-Mobile understands that public safety continues to be concerned
about the provision of more accurate indoor (and outdoor) location estimate information—as is
T-Mobile. T-Mobile is working both on its own and as part of the FCC’s Communications,
Security, Reliability and Interoperability Council (“CSRIC”) and the Alliance for
Telecommunications Industry Solutions’ (“ATIS”) Emergency Services Interconnection Forum
(“ESIF”) to explore new technological solutions that promise to narrow the search radius that TMobile reports. This includes moving to integrate into handsets additional satellite constellations
(i.e., GLONASS, another Global Navigation Satellite System (“GNSS”)) to complement the
Global Positioning System (“GPS”) constellation, and Observed Time Difference of Arrival
(“OTDOA”) to improve location accuracy when GNSS systems cannot provide a high accuracy
location estimate on their own. Other technologies of interest are maturing to the point of
nearing readiness for consideration, such as Wi-Fi-based location methods, currently used
exclusively for commercial location-based services (“cLBS”). Furthermore, when T-Mobile
implements voice-over-Long Term Evolution (“VoLTE”), it will gain the ability to run multiple
location technologies simultaneously, rather than sequentially, improving its ability to deliver
highly accurate location estimates within a given latency limit.

2

 When implemented, these measures will improve what is already a high level of
performance. Wireless 911 and E911 are working well to enable people to place millions of 911
calls every year. Today, T-Mobile is able to provide a Phase II location estimate for at least 90
percent of wireless 911 calls made on its network—including calls made from indoors. If threequarters of wireless 911 calls are made from indoors, as some have speculated, and T-Mobile is
able to provide an Assisted GPS (“A-GPS”) fix for 77 percent of E911 calls from its subscribers,
it follows that at least two-thirds of all of T-Mobile’s indoor wireless 911 calls of more than 30
seconds in length are getting position fixes based on A-GPS.3 And T-Mobile’s A-GPS yield and
accuracy performance continues to improve over time. Moreover, T-Mobile delivers uncertainty
estimates at a 90 percent confidence level—the level public safety requested through ESIF bestpractice recommendations—to any Public Safety Answering Point (“PSAP”) that wants it. That
uncertainty and confidence information allows a PSAP to ascertain—irrespective of underlying
location technology—whether it is getting an estimate that is likely to be highly accurate, or one
with a broader potential range of error.
Even when these emerging technologies are fully implemented, though, they will not
enable carriers to meet the Commission’s proposed new horizontal indoor accuracy requirements
of 50 meters for 67 percent of calls within two years and 50 meters for 80 percent of calls within
five years. Once such solutions are market-ready—which is not presently the case—they will
need to be incorporated into standards and deployed throughout the network as well as, for any
handset-based solutions, into handsets. Even if these solutions were ready for commercial
deployment today, unless the Commission is going to require every wireless consumer to turn in
their phones and buy new ones, it will take years for consumers to exchange their old handsets
3

Letter from John T. Nakahata, Counsel to T-Mobile USA, Inc., to Marlene H. Dortch,
Secretary, FCC, PS Docket No. 07-114 (filed Nov. 18, 2013).
3

 for new ones with additional location capabilities.4 How many years this subscriber handset
change-out process takes is impossible to accurately predict and is only partially influenced by
wireless carriers. T-Mobile notes that the handset change-out during the initial implementation
of Phase II took far longer than expected; that experience should inform any realistic evaluation
of the time frame for consumer adoption of new devices.
This is also true with respect to the Commission’s proposed vertical indoor location
accuracy requirement of 3 meters for 67 percent of calls within three years, and 3 meters for 80
percent of calls within five years. If barometric pressure sensors turn out to be necessary for the
provision of vertical location estimates, a wireless carrier would only be able to attempt to
comply with such a mandate after its customers change out their current handsets for ones with
barometric pressure sensors, as very few handset models currently have pressure sensor
capability. But the use of barometric pressure sensors to measure altitude has only begun to be
tested rigorously, and already studies indicate that it is highly unlikely that these sensors will be
capable of providing accurate estimates of actual altitude, absent the placement of reference
sensors in every multistory building.
The Administrative Procedure Act does not permit the Commission to take an “if we
mandate it, they will invent it” approach. Technical and economic feasibility are “made
necessary by the bar against arbitrary and capricious decision-making,”5 and “[i]mpossible

4

Indeed, the Commission’s recent announcement of the voluntary commitment among AT&T,
Sprint, T-Mobile, U.S. Cellular, and Verizon Wireless regarding handset unlocking is likely
to lead to greater handset retention by consumers, potentially making the handset exchange
cycle longer. See Letter from Steve Largent, President/CEO, CTIA, to Thomas E. Wheeler,
Chairman, FCC, et al (Dec. 12, 2013), available at http://hraunfoss.fcc.gov/edocs_public/
attachmatch/DOC-324664A1.pdf.

5

Nuvio Corp. v. FCC, 473 F.3d 302, 303 (D.C. Cir. 2006).
4

 requirements imposed by an agency are perforce unreasonable.”6 Congress has not given the
Commission the authority to promulgate rules that force carriers to develop and adopt
technologies that are not proven to be feasible, and for which there is no record evidence of that
feasibility.7 The Commission should, in the near term, call for the establishment of an expedited
test bed to explore location technologies of high interest, such as OTDOA, the use of additional
satellite constellations, and even Wi-Fi-based location technologies, following the model that
CSRIC III used. In the longer term, the Commission should also support the creation of a
permanent indoor test bed to test new and emerging location technologies. This will allow both
a common evaluation of proposed technologies as well as a feasible means for ascertaining
compliance.
No matter what, it will take many years—more than the five contemplated in the
Commission’s proposed rules—to develop and implement enhanced indoor location technologies
in consumers’ handsets. And even with the Commission’s proposed new requirements, at the
end of that process public safety will still not have an actionable address, including floor level,
for dense urban environments. Rather than just trying to shrink the location estimate circle by a
number of meters, it would be much better to end this multiyear transition with an indoor
location solution that really meets public safety’s needs and delivers actionable dispatch
information. But this can only be accomplished by forging a more creative and effective multistakeholder approach, one that looks not only to carriers, technology vendors, mobile operating

6

Alliance for Cannabis Therapeutics v. Drug Enforcement Admin., 930 F.2d 936, 940 (D.C.
Cir. 1991).

7

Cf. Edison Elec. Inst. v. EPA, 996 F.2d 326, 337 (D.C. Cir. 1993) (“[T]he fact that
technology may not be able to keep up with timetables established by Congress does not
mean that courts are at liberty to ignore them, however burdensome the resulting
enforcement.”). Congress has enacted no such legislation here.
5

 systems providers, and public safety, but also to state and local governments as well as premises
owners.
II.

LOCATION ACCURACY IS BEST IMPROVED THROUGH SUPPORT OF INDUSTRY EFFORTS
AND MULTI-STAKEHOLDER TESTING.
A.

Current and Near-Term Technologies Will Continue to Improve Location
Estimates for Wireless 911 Calls.

Wireless E911 continues to improve as carriers and PSAPs work to ensure that wireless
calls to 911 are accompanied by the most accurate information possible. This is borne out by
reality—for the vast majority of 911 calls, the caller is able to summon the help he or she needs.
Much of the ongoing improvement in wireless location accuracy and yield is due to
carrier efforts. As T-Mobile described in 2011, some of these efforts include improving network
and handset algorithms and handset GPS receiver sensitivities. As T-Mobile has implemented
improvements in its network and handsets, it has seen its A-GPS yield steadily increase—for
instance, from 2012 to 2013, it increased from 74 to 77 percent. T-Mobile expects to see
continued improvement from ongoing updates to its existing A-GPS deployment.
Improvements in the effectiveness of wireless location information have also resulted
from changes in the way PSAPs process 911 calls. For instance, after CalNENA submitted
information to the Commission last summer suggesting a marked drop-off in Phase II
information,8 it became clear that the relevant PSAPs—and many other PSAPs nationwide—
were not following the Association of Public Safety Communications Officials International
(“APCO”) best practices and rebidding for updated location information.9 Those PSAPs that

8

See Letter from Danita L. Crombach, ENP, CalNENA, to Mignon Clyburn, Chairwoman,
FCC, PS Docket No. 07-114, at 3-4 (filed Aug. 12, 2013).

9

APCO Project LOCATE, Wireless 9-1-1 Deployment and Management Effective Practices
Guide, Effective Practice 380743 (2007), available at https://dps.mn.gov/divisions/ecn/
programs/911/Documents/APCO_LOCATE_Effective_Practices.pdf; Letter from Steve
6

 have adopted policies calling for rebidding at appropriate intervals—or that follow recommended
processes for automatic rebidding—see higher Phase II yields at the PSAP10 and a subsequent
increase in perceived location accuracy.11
Additional improvement to wireless location is on the horizon, as well. New satellite
constellations (i.e., GLONASS) are being integrated into carrier networks, and carriers are
working to ensure that their handsets can communicate with those satellites, which will allow
them to provide more accurate and higher yield satellite-based location fixes. Other
technologies, like OTDOA, are coming to market and will be deployed soon. Indeed, T-Mobile
has committed financial and personnel resources to implementation of OTDOA throughout its
network and is currently actively working to deploy it, even without a mandate. And carriers
continue to explore the possibility of leveraging commercial location services that are already
available on many handsets. Each of these emerging location technologies hold promise to
further improve automatic location information, including indoors—but each needs careful
evaluation and testing through the CSRIC test bed methodology before specific performance
standards are established.

Largent, President/CEO, CTIA, to Tom Wheeler, Chairman, FCC, PS Docket No. 07-114
(filed Dec. 5, 2013) (“Dec. 5, 2013 CTIA Letter”).
10

See Dec. 5, 2013 CTIA Letter; Letter from Marlys Davis, King County E-9-1-1 Program
Office, to Marlene H. Dortch, Secretary, FCC, PS Docket No. 07-114 (filed Sept. 25, 2013);
Letter from Stan Heffernan, Chief Operations Officer, Greater Harris County Emergency
Network, to David Siehl, Public Safety and Homeland Security Bureau, FCC (Sept. 16,
2013), available at http://transition.fcc.gov/bureaus/pshs/911/Phase%202/TX/
Greater%20Harris%20Co%20TX%20FCC%20WPH2%20Letter%20091613.pdf.

11

T-Mobile reiterates that it develops Phase II location information for the vast majority of
calls of sufficient duration to allow a location fix to be calculated. The PSAP must request
that location fix, however, and if it does not, the PSAP may perceive a lower Phase II yield
than is actually available. Current signaling standards do not allow a wireless carrier to
“push” location information to the PSAP.
7

 B.

The Commission Should Move to Establish a Permanent Indoor Test Bed
and Rely on CSRIC’s Recommendations Regarding New and Emerging
Location Technologies.

T-Mobile and other stakeholders have repeatedly called for the establishment of a
permanent indoor test bed that will allow for investigation of new and emerging location
technologies.12 It is critical that all stakeholders have the opportunity to evaluate technologies
that have the potential to improve wireless E911 location accuracy in a common, standardized,
independently-run environment. In other words, the use of a single test bed to which all location
vendors submit their technologies for evaluation will provide a real apples-to-apples comparison
in regards to truly viable solutions. T-Mobile is willing to shoulder its share of supporting the
expedited creation and operation of such a test bed, provided that other national carriers and
relevant stakeholders do likewise.
In the near term, T-Mobile also supports expediting the establishment of a targeted test
bed to evaluate technologies of current high interest—specifically, those technologies that were
not available for testing in the CSRIC III test bed.13 Those technologies include emerging LTE
technologies like OTDOA, which holds a great deal of promise, as well as the newly available
GLONASS satellite constellation that, when integrated into handsets, is expected to complement
A-GPS, thus improving location accuracy and yield. These important technological
advancements are already standardized and committed for deployment by national carriers but—

12

Comments of T-Mobile USA, Inc., PS Docket No. 07-114, at 27-28 (Sept. 25, 2013) (“TMobile Comments on Location Accuracy Workshop”).

13

See The Communications Security, Reliability, and Interoperability Council III, Working
Group 3, Indoor Location Test Bed Report at 55 (2013) (“CSRIC III WG3 Report”) (noting
four technologies that expressed interest but did not participate in the Stage 1 test bed, and
three additional technologies that were not available in the Stage 1 timeframe but that are
good candidates for the next stage).
8

 very importantly—have yet to be evaluated in a common test bed to determine real-world
performance in a variety of indoor environments.
The addition of Wi-Fi-based location information currently utilized for cLBS is also an
important technology that carriers and other stakeholders may be able to leverage to improve
E911 location accuracy. These technologies should also be tested in a common, independently
run test bed, allowing stakeholders as well as the Commission to determine what truly is
technically and economically feasible.
T-Mobile believes that a test bed targeted for these specific technologies of high interest
could be established in an expeditious manner, with the full backing and support of the
Commission, other wireless carriers, public safety, and other relevant stakeholders (including
location technology vendors, mobile operating system providers, and others). The methodology
that was developed and successfully implemented by CSRIC III could be followed, with
oversight from a technical advisory group made up of the various stakeholders involved in the
process (as was done for the CSRIC III test bed effort).
The establishment of a permanent test bed will further allow for ongoing testing of new
and emerging technologies, rather than providing only a snapshot in time of available
technologies. The ability to compare a new candidate technology against existing technologies
in a standardized environment will allow stakeholders to develop reasonable expectations for
near- and long-term location accuracy. This will enable carriers, handset manufacturers,
vendors, public safety, and the Commission to make informed choices.
III.

TECHNICAL AND ECONOMIC FEASIBILITY AS WELL AS PUBLIC SAFETY’S NEED FOR
“TRULY ACTIONABLE” LOCATION INFORMATION MUST GUIDE THE COMMISSION’S
DECISIONMAKING.
Any new rules adopted for indoor location accuracy must be firmly grounded in technical

realities. T-Mobile urges the Commission, therefore, to refrain from imposing new indoor
9

 location accuracy standards until the establishment of an indoor test bed to evaluate emerging
location technologies. Only such standardized, common-environment testing can make clear
what level of location accuracy performance is feasible. If the Commission chooses to move
forward with new rules before the new test bed is established, it must ensure that those rules
acknowledge the limitations of the available technologies that were made clear in the CSRIC III
test bed.
The currently proposed 50 meter horizontal and 3 meter vertical accuracy benchmarks
ignore the CSRIC III test bed results, which clearly demonstrate the difficulties of currently
available technologies to meet the proposed benchmarks in challenging environments. In the
face of those results, the Commission’s proposed rules are simply not achievable—certainly not
within the Commission’s proposed time frames—particularly given that the Commission has
based these benchmarks in part on the test results of a single beacon-based technology that has
only been deployed in one metropolitan area and cannot be accessed by any currently or nearterm available handset.
Furthermore, the Commission must not forget that, even if carriers can provide such
improved location information, it will always be delivered in the form of x, y and, for vertical
estimates, z coordinates. Such three-dimensional indoor location information is likely to be
useless to PSAPs until they upgrade their systems and mapping data to include detailed inbuilding elevation and floor layout data. Without those upgrades, improved location
information—and particularly vertical information—will not benefit a single 911 caller.
Buildings vary in the height of their floors, and some floors (such as the first floor) may be
significantly taller than others. Accordingly, a wireless carrier cannot tell the PSAP what floor a
caller is on simply based on the height above mean sea level, regardless of the underlying

10

 accuracy of the vertical location estimate. And the PSAP upgrades required to translate an
estimate of height above mean sea level into an actual estimated floor number are likely to take
far longer than the five years proposed by the Commission. In this, the Commission should
recall the lengthy implementation period required for widespread adoption of Phase II
technology in PSAPs and recognize that PSAPs will continue to be limited in their ability to
quickly deploy the necessary upgrades for acceptance and use of vertical location information.
A.

The Indoor Location Proposals Ignore the Results of the CSRIC Test Bed.

No technology or combination of technologies deployed by carriers can reach the
Commission’s proposed new indoor location accuracy benchmarks, or provide a location so
specific that first responders know “what door to kick in.”14 Though the CSRIC test bed
suggests that there are candidate technologies that show promise for improved indoor location
accuracy, none of the technologies tested were capable of achieving the benchmarks proposed by
the Commission. The Test Bed Report expressly found that “[w]hile the location positioning
platforms tested provided a relatively high level of yield, as well as improved accuracy
performance, the results clearly indicate additional development is required to ensure the
positional coordinates provided on an emergency caller sheltered indoors result in an ‘actionable
location’ for emergency response, especially in urban and dense urban environments.”15
Indeed, in the CSRIC test bed, even the best-performing technologies could not reliably place
calls in the right buildings, with only about a third of the fixes provided by the most accurate
location technologies in the test bed falling inside the target buildings.16 These issues are likely
14

Wireless carriers, of course, will never be able to tell PSAPs “what door to kick in”—they
can only provide geographic coordinates. PSAPs are, and continue to be, responsible for
converting those coordinates to actionable addresses.

15

CSRIC III WG3 Report at 8.

16

Id. at 39.
11

 to be even more pronounced where buildings are closely spaced, as in urban and dense urban
environments.
1. NextNav’s Beacon Technology Is No Panacea.
The Commission cites NextNav’s beacon-based technology as evidence that there are
existing location technologies that can be implemented and that will allow carriers to meet the
proposed indoor location accuracy benchmarks within the proposed time frame.17 But this
conclusion ignores several problems with NextNav’s technology.
First, NextNav does not purport to offer service on a nationwide basis. Its proposal has
always been to deploy its beacon network in select markets.18 In fact, NextNav does not have
the spectrum licenses needed to deploy and operate its technology outside of some quantity of
the top metro areas.19 Furthermore, to date, NextNav has only deployed in one area to the
density it believes is required to provide accurate location estimates—the area where the CSRIC
tests were performed. Deployments in the other areas where NextNav holds licenses are just
beginning.20 The Commission should be wary of adopting rules based on a technology for which
no nationwide deployment is planned.
Second, there is significant reason—notwithstanding the results of the CSRIC Working
Group III tests—to question whether NextNav in fact has deployed, or will be able to deploy, a
sufficient number of beacons to sustain highly accurate vertical location estimates. Initial studies
indicate that barometric sensors—which are the foundation for the vertical estimates in

17

Third FNPRM at 2394, 2401-03 ¶¶ 47, 71-74.

18

Letter from Bruce A. Olcott, Counsel to Progeny LMS, LLC, to Marlene H. Dortch,
Secretary, FCC, WT Docket No. 11-49, Attachment at 13 (filed Mar. 21, 2013).

19

Id.

20

Id.
12

 NextNav’s or any similar system—will face significant challenges in reliably determining
absolute vertical location in many different types of buildings and environmental conditions.
Critically, though the use of barometric pressure to track relative changes in altitude over
short periods of time is relatively straightforward, a barometric pressure measurement by a
handset alone cannot produce an accurate absolute altitude estimate either indoors or outdoors.21
Accurate absolute vertical measurements must rely on an initial pressure reference (calibration)
measurement made in real-time from a sensor at a known altitude in the same atmospheric
conditions as the handset. While it may be reasonable in some cases to assume that an outdoor
calibration sensor will be in the same atmospheric conditions as a handset inside a building, in
many other cases, this assumption will not be reasonable. Some challenges arise from
mechanical systems common in semi-sealed, environmentally controlled buildings (which
includes most modern multistory office buildings). In these buildings, fan pressurization alone
can typically increase or reduce in-building pressure by as much as 50Pa (0.5hPa) relative to
outdoor static air pressure.22 This impact can affect measured altitude by more than 4 meters23—
even with highly accurate, local, outdoor pressure calibration in place. Other challenges arise in
older, “leaky” (high infiltration rate) buildings, where natural processes such as the “stack
effect,” rapidly changing weather conditions, cold temperatures, high wind conditions, etc. can
21

See generally Binghao Li et al., Using Barometers to Determine the Height for Indoor
Positioning, International Global Navigation Satellite Systems Society (July 2013), available
at http://www.gmat.unsw.edu.au/snap/publications/lib_etal2013a.pdf; see also Kartik
Muralidharan et al., Barometric Phone Sensors – More Hype Than Hope!, ACM HotMobile
(Feb. 2014), available at http://research.microsoft.com/apps/pubs/default.aspx?id=206442.

22

Rick Quirouette, Air Pressure and the Building Envelope, at 6 (Nov. 2004), available at
https://www.cmhc-schl.gc.ca/en/inpr/bude/himu/coedar/upload/Air-Pressure-and-theBuilding-Envelope.pdf (“Air Pressure and the Building Envelope”).

23

Using the “standard atmosphere” pressure model, pressure drops by about 0.11hPa for every
1 meter increase in altitude. Alternatively, altitude changes by about 8.7 meters for every
1hPa change in air pressure. (1hPa = 100Pa)
13

 greatly affect internal pressure measurements relative to outdoor static air pressure. The stack
effect alone in cold temperature conditions can create a pressure gradient in the building air
column and a resulting indoor-to-outdoor pressure difference of approximately 90Pa (0.9hPa) on
the top floor of a 20-story building, even in the absence of wind.24 This pressure difference
would introduce an altitude measurement error of nearly 8 meters, even with highly accurate,
local outdoor pressure calibration. These indoor-to-outdoor pressure differences and resulting
altitude measurement errors can be much larger for taller buildings and/or high wind conditions.
While the most apparent solution would be to require the installation of barometric pressure
sensors inside each building to be used as references for that building, and at various heights for
tall buildings, such a requirement is outside the purview of wireless service providers, making a
ubiquitous requirement for carriers to provide accurate in-building altitude measurements
infeasible. And, further, if a solution involves installation of new devices in buildings, it is not at
all clear that barometric pressure sensors are the best solution. Other technologies may be more
effective.25
In addition to the inaccuracies arising from differences between the atmospheric
conditions inside and outside of buildings, there will always be errors in the barometric pressure
measurement by the handset. Sensor errors fall into two categories: the relative measurement
error and absolute error, or “sensor drift.” Relative measurement error for these sensors is

24

Air Pressure and the Building Envelope at 5.

25

One possible example would be the use of in-building Bluetooth-based location beacons.
The deployment and use of this type of device could be required through building codes in
much the same way that codes currently require the deployment of smoke detectors, fire
extinguishers, or exit signs.
14

 typically rated at no better than plus or minus one meter.26 This error source cannot be calibrated
out of the resulting altitude measurement and, under the proposed vertical benchmark, this would
leave a “calibration error budget” of only two meters to ensure that a given measurement falls
within the three meter overall accuracy requirement. “Sensor drift,” meanwhile, causes sensor
measurement to drift from 60 to 105 meters over time and temperature. This error source can be
calibrated out of the measurement, but only if an accurate and frequent reference measurement is
available at a known altitude in the same air column environment as the handset. As discussed
above, it remains unclear (if not unlikely) whether such accurate reference measurements will be
available.
These challenges are mostly unexplored in the E911 context, as the CSRIC III test bed
was not designed to test altitude measurements under the conditions that are most challenging for
barometric pressure sensor methodology. That test bed primarily focused on evaluating
horizontal location performance. T-Mobile urges the Commission to ensure that vertical location
performance of this technology—or any similar technology—is thoroughly tested, evaluated, and
standardized in all appropriate building and environmental conditions before adopting any new
rules based on its potential capabilities. Mandating inclusion of unproven, costly technology into
all handsets (and infrastructure) will, ultimately, be burdensome to consumers and may delay the
development of more effective technologies in the future.
Third, NextNav’s terrestrial beacon technology did not prove capable of meeting the
proposed 50 meter horizontal accuracy standard in all indoor environments in the CSRIC III test

26

See, e.g., Bosch, BMP180 Digital, barometric pressure sensor, Specification Sheet, Doc. No.
BST-BMP180-FL000-02, available at http://ae-bst.resource.bosch.com/media/downloads/
pressure/bmp180/Flyer_BMP180_08_2013_web.pdf.
15

 bed at either the 67th or 80th percentiles.27 Specifically, both the urban and dense urban
environments (precisely the areas where AGPS technology is most challenged) fell well outside
the 50 meter proposed requirement.28 NextNav also fell outside the proposed 3 meter vertical
requirement in the suburban environment.29
Fourth, setting aside these fundamental technical issues, there are no handsets on the
market, and none in the pipeline, that have built-in NextNav terrestrial beacon technology.
Indeed, NextNav is just beginning the process of working with standards bodies to determine
how its technology can be accommodated within existing mobile technology frameworks.30 For
the CSRIC tests, NextNav utilized prototype hardware external to the mobile phone—a
“sleeve”—which may not be representative of performance achievable in a production handset
form factor. Handset integration is a key limiting factor in deploying even GPS-based location
technology;31 it will undoubtedly be an even greater hurdle for third-party location technologies
like NextNav’s, where the technology must be introduced into the mobile ecosystem—from
standards to devices—from scratch.32

27

See CSRIC III WG3 Report at 27.

28

In the urban setting, NextNav’s horizontal location technology demonstrated 63 meters at 67
percent and 85 meters at 80 percent. In the dense urban setting, NextNav demonstrated 57
meters at 67 percent and 73 meters at 80 percent. Id. at §§ 7.3.1-2.

29

In the suburban setting, NextNav’s vertical location results were 4.5 meters at 67 percent and
5 meters at 80 percent. Id. at 37, Figure 7.5-1.

30

Letter from Bruce A. Olcott, Counsel to NextNav, LLC, to Marlene H. Dortch, Secretary,
FCC, PS Docket No. 07-114 (Apr. 23, 2014) (“NextNav Ex Parte”).

31

See T-Mobile USA, Inc. Comments On Section III.B of the Wireless E911 Location
Accuracy NPRM, PS Docket No. 07-114 & WC Docket No. 05-196, at 19 (Aug. 20, 2007)
(“T-Mobile 2007 Accuracy NPRM Part B Comments”).

32

See NextNav Ex Parte (noting that NextNav is working with standards-setting bodies to,
among other things, develop a standard to address air interface and control and user plane
messaging, as well as working with carriers and manufacturers to incorporate NextNav
technology into chipsets and devices).
16

 2. Other Location Technologies Are Similarly Challenged.
The Commission acknowledges that the “test bed results indicate that further
improvement is necessary,”33 but nevertheless proposes adopting a 50 meter search ring in the
belief that such accuracies are “attainable in the near-term.”34 In support of this proposal, it cites
the test bed results for NextNav’s beacon-based system—the flaws of which are discussed
above—as well as technologies offered by Polaris (Radio Frequency (“RF”) Pattern Matching),
Qualcomm (AGPS-Advanced Forward Link Trilateration (“AFLT”)), and TruePosition (AGPSUplink-Time Difference of Arrival (“U-TDOA”)). To the extent the Commission is relying on
test results of any of these technologies to support the 50 meter indoor benchmark, that reliance
is misplaced.
First, the Test Bed Report makes clear that neither Polaris’s nor Qualcomm’s
technologies are currently capable of meeting a 50 meter accuracy requirement for either 67
percent or 80 percent of calls in almost any indoor environment.35
Second, TruePosition’s technology has proven time and again to be incapable of
providing highly accurate location estimates. U-TDOA has also proven economically infeasible
to deploy and maintain in comparison to other, better performing technology alternatives.
Indeed, as T-Mobile has noted on the record in this docket, carriers have migrated away from UTDOA because of its inherent shortcomings, including poor performance in rural environments
and environments such as highways, beaches, edges of service areas, and other areas with

33

Third FNPRM at 2394 ¶ 47.

34

Id. at 2393 ¶ 45.

35

CSRIC III WG3 Report at § 7.3. The sole exception is that Qualcomm demonstrated 48.5m
accuracy at 67% in the rural indoor environment. Other results for these two technologies in
the urban and dense urban environments ranged from 117 meters to 227 meters at the 67th
percentile, and 190 meters to 306 meters at the 80th percentile.
17

 suboptimal cell geometry.36 Furthermore, TruePosition chose not to participate in the CSRIC
test bed—rather, it conducted its own testing, in a very different environment, without the benefit
of direct comparison with other technologies, and using a very different reporting
methodology.37 The Commission should not rely on those results to support the 50 meter indoor
benchmark; TruePosition’s U-TDOA technology needs further apples-to-apples testing in a
standardized test environment.38
B.

The Indoor Location Proposals Establish Unrealistic Time Frames Given the
Need to Test and Deploy Equipment as well as for Handset Turnover.

The Commission has stated that “the adoption of more stringent requirements for indoor
location accuracy, together with a reasonable implementation timeframe, would afford CMRS
providers with sufficient time and incentive to develop the necessary technology to enable
compliance.”39 Even assuming that technological solutions that were capable of meeting a 50
meter indoor horizontal requirement and a 3 meter vertical accuracy requirement for 67 percent
of calls were available in commercial-ready prototypes today—which is not the case—the
proposed timelines cannot possibly be met.
As T-Mobile has noted on the record in the past, “meeting any new accuracy requirement
that is more stringent than the current requirements, or even implementing the current
requirements at the PSAP level, will require installation of new technology and swapping out
existing technology,” requiring an extended transition period.40 Such transitions require years,

36

See T-Mobile Comments on Location Accuracy Workshop at 29-31.

37

See id. at 30 (citing TruePosition test bed report).

38

See id. at 28-30.

39

Third FNPRM at 2394 ¶ 47.

40

T-Mobile 2007 Accuracy NPRM Part B Comments at 17.
18

 not months, to ensure that they are incorporated into industry standards as well as into actual
equipment, software, and products.41
Any new technologies must first be incorporated into industry standards, a process that
takes at least 12 to 24 months in the appropriate standards bodies. This ensures that the
technologies work and properly interoperate with the rest of the commercial mobile radio service
(“CMRS”) world. It also helps to ensure that a given technology can be developed and offered
by multiple vendors, ensuring a healthy ecosystem and continued innovation.
Only after new standards are developed can they be implemented into equipment and
software; indeed, most telecommunications manufacturers will not even consider the inclusion of
non-standardized technologies into their devices and network equipment. The design,
development, and implementation time can take two or three years, as the U.S. market is only a
part of global CMRS equipment markets and therefore must be coordinated with other product
changes being implemented around the world. And deployment of new network equipment
throughout networks itself takes time—at least two years, perhaps longer if tower loads are
increased or additional ground space is needed.
Perhaps most critically, new technologies must be built into handsets, which must then be
replaced across the subscriber base. Though handset turnover times may arguably be shorter
now than they were a decade ago, it remains a lengthy process, and it is difficult to accelerate
without requiring consumers to replace handsets and/or mandating that carriers shut off service
to older handsets. The Commission can therefore expect handset turnover alone to take several
years. And because next-generation location accuracy technologies universally will require new
41

While T-Mobile, as a carrier, is best able to speak to the transition timing for carrier and
subscriber equipment, as discussed above, even if carriers can provide vertical location
information, it will be useless until PSAPs upgrade their systems and mapping data to
include detailed building elevation and floor layout data.
19

 handsets, carriers will not be able to meet the proposed accuracy percentages until new handset
penetration reaches a critical point, presumably at least 67 percent of subscribers for the nearterm requirement and at least 80 percent for the long-term requirement.
MANY OF THE COMMISSION’S PROPOSALS APPEAR TO CONFLICT WITH ONE ANOTHER
AND CREATE INCONSISTENT OR CONFUSING OBLIGATIONS.

IV.

The Commission should clarify a number of its proposals, as described in more detail
below.
First, the Commission has proposed what appear to be two conflicting compliance
regimes—one based on county or PSAP level results, and one based on testing in a
representative test bed that would establish nationwide compliance. T-Mobile continues to
believe that the Commission must establish a permanent indoor location accuracy test bed, and
that relying on test bed data for compliance is far preferable to attempting to measure compliance
on a county or PSAP basis.42 In fact, logistics alone dictate against measuring on a county or
PSAP basis, as it would take unreasonable amounts of time to negotiate building access with
landlords and establish ground truth positions for each test location. Furthermore, once a
technology or combination of technologies has demonstrated compliance through the test bed
process, ongoing, periodic compliance testing should not be required. Carriers introducing a new
location methodology should be permitted to demonstrate initial compliance via the compliance
test bed process, and the passage of time alone should not require automatic re-testing.
Second, if the Commission establishes a maximum time to first fix of 30 seconds43—
which T-Mobile supports—then calls of less than 30 seconds’ duration (rather than calls of less

42

See Comments of T-Mobile USA, Inc., PS Docket No. 07-114, at 23 (Jan. 19, 2011).

43

Third FNPRM at 2430-31 ¶ 148.
20

 than 10 seconds44) should be excluded. Where carriers are permitted up to 30 seconds to
establish a Phase II location fix, a rule excluding from consideration only calls of less than 10
seconds will create confusion and uncertainty. The Commission should harmonize these two
proposals by excluding calls of less than 30 seconds’ duration.
Third, and finally, requiring carriers to provide PSAPs with the specific location
technology used to generate a location estimate is unlikely to accomplish anything other than to
confuse call takers. Call takers do not need to see an alphabet soup of various location
technologies when answering emergency calls. They only need to know the search radius, and
that information is contained in the uncertainty measurements at a given confidence level. The
FNPRM correctly identifies that uncertainty information is less useful to PSAPs than it might be
due to the fact that carriers currently provide uncertainty at varying confidence levels. That
problem, however, could be eliminated by adopting a standardized confidence level for
uncertainty measurements. T-Mobile supports the Commission’s proposal to establish an
industry standard confidence level of 90 percent for uncertainty estimates.45 Doing so should
obviate the need to provide PSAPs with the specific technology or technologies used to estimate
a caller’s location.
V.

THE COMMISSION SHOULD REFRAIN FROM MODIFYING THE EXISTING LOCATION
ACCURACY FRAMEWORK FOR OUTDOOR CALLS.
The outdoor location accuracy rules function well, and carriers are on a path to a unified

accuracy standard. The percentage of PSAPs in which a carrier is required to meet the accuracy

44

Third FNPRM at 2430 ¶ 147.

45

T-Mobile recognizes that some carriers utilize a handset-based location architecture where
the position calculation function is performed entirely in the handset, and may be unable to
modify the confidence level for existing handsets. In this case, a requirement to adopt
standardized confidence and uncertainty could be limited to those handsets introduced
beyond a certain date.
21

 standards continues to ratchet up.46 This path remains very challenging, and compliance—even
with today’s best performing technologies—is far from certain. Rather than accelerate the
outdoor location benchmarks or adopt a unitary standard on a short time frame, the Commission
should retain the existing benchmarks. As T-Mobile has said before, the move to handset-based
accuracy will always be subject to the ability of carriers to roll out new handsets throughout their
footprint. That process is difficult if not impossible to accelerate. Pushing the existing time
frame up would impose enormous burdens on carriers, at precisely the same time that those
carriers are working hard to deploy new network technologies that hold the promise of real
improvements to location accuracy in all environments.

46

See 47 C.F.R. § 20.18.
22

 VI.

CONCLUSION
T-Mobile continues to work toward improved location accuracy for wireless E911,

including for indoor wireless calls. But in the face of clear evidence that no existing or near-term
technology can meet the Commission’s proposed indoor accuracy benchmark, it urges the
Commission to refrain from adopting new rules and instead support the establishment of a
permanent indoor test bed. A multi-stakeholder, collaborative approach is needed to ensure the
success of any new technologies eventually adopted, as well as to enable the Commission to
realize its stated goal of ensuring that PSAPs have a “truly actionable” location for all wireless
E911 calls.
Respectfully submitted,

Steve B. Sharkey
Eric Hagerson
T-MOBILE USA, INC.
601 Pennsylvania Ave., NW
Washington, DC 20004
(202) 654-5900

John T. Nakahata
Kristine Laudadio Devine
HARRIS, WILTSHIRE & GRANNIS LLP
1200 Eighteenth Street, NW
Suite 1200
Washington, DC 20036
(202) 730-1300
Counsel to T-Mobile USA, Inc.

May 12, 2014

23