Skyhook Wireless, Inc. v. GOOGLE, INC.

Filing 43

DECLARATION of Anthony S. Acampora, Ph.D. by GOOGLE, INC.. (Attachments: # 1 Exhibit 1)(Manning, Susan)

UNITED STATES DISTRICT COURT
DISTRICT OF MASSACHUSETTS

SKYHOOK WIRELESS, INC.,
Plaintiff,
v.
GOOGLE INC.,
Defendant.

GOOGLE INC.,
Counterclaim-Plaintiff,
v.
SKYHOOK WIRELESS, INC.,
Counterclaim-Defendant.

)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)

CIVIL ACTION
NO. 1:10-cv-11571-RWZ

DECLARATION OF ANTHONY S. ACAMPORA., PH.D.

A/74514379.11/3005005-0000352152

TABLE OF CONTENTS
Page
I.
II.

III.
IV.

V.

QUALIFICATIONS AND ENGAGEMENT.................................................................... 1
THE PATENTS-IN-SUIT ................................................................................................. 6
A.
Background and Description of the Claimed Invention. ....................................... 6
B.
The asserted claims. ............................................................................................... 8
C.
The asserted patents’ familial relationship to each other..................................... 11
D.
The alleged invention of the asserted claims—”Arterial Bias,” “Reference
Symmetry,” and “Chinese Postman.” .................................................................. 12
E.
The Prosecution History of the ‘988 Patent......................................................... 15
F.
The Prosecution History of the ‘694 Patent......................................................... 20
G.
The Prosecution Histories of the ‘245 and ‘897 Patents...................................... 22
SUMMARY OF STIPULATED AND DISPUTED CLAIM CONSTRUCTIONS........ 23
CLAIM CONSTRUCTION PRINCIPLES ..................................................................... 26
A.
The Level of Ordinary Skill in the Art................................................................. 26
B.
Specification, Prosecution History....................................................................... 26
C.
Means-plus-function claiming. ............................................................................ 27
D.
Indefiniteness ....................................................................................................... 28
SEVERAL TERMS IN THE PATENTS-IN-SUIT ARE INDEFINITE......................... 29
A.
The “Reference Symmetry” Limitations are Indefinite. ..................................... 29
B.
“Arterial bias” and “avoid(s) arterial bias.”......................................................... 31
C.
The “Logic” Terms. ............................................................................................. 33
1.
The “logic” terms are functional.............................................................. 34
2.

The claims and specification of the ‘988 patent do not disclose any
corresponding structures for the “logic” terms. ....................................... 35
a.
“logic to recalculate position information for Wi-Fi access
points previously stored in the database to utilize position
information for the newly-discovered readings of
previously stored Wi-Fi access points” (claim 1) ........................ 35
b.
“computer-implemented logic to add records to the
database for newly-discovered Wi-Fi access points” (claim
1) .................................................................................................. 39
c.
“computer-implemented clustering logic to identify
position information based on error prone GPS
information” (claim 2) ................................................................. 41
d.
“the clustering logic … excludes such deviating position
information from the database and from influencing the
calculated positions of the Wi-Fi access points” (claim 3).......... 44

i
A/74514379.11/3005005- 0000352152

TABLE OF CONTENTS
(continued)
Page
e.

VI.

“logic to determine a weighted centroid position for all
position information reported for an access point” (claim 3) ...... 44
f.
“logic to identify position information that exceeds a
statistically-based deviation threshold amount away from
the centroid position” (claim 3) ................................................... 47
D.
“Predefined Rules” and “Rules” (‘897 patent, claims 1 and 3) ........................... 49
E.
“Being suited” (‘245 patent, claim 1) .................................................................. 50
THE PROPER CONSTRUCTION OF CERTAIN DISPUTED CLAIM TERMS......... 50
A.
“Target area”....................................................................................................... 50
B.
The Location Terms: “calculated position information” (‘988/1 & ‘694/1);
“calculated positions of the Wi-Fi access points” (‘988/3); “calculated
locations” (‘245/1 & 2); “recorded location information” (‘897/1 & 3) ............. 52
C.
“substantially all Wi-Fi access points in the target area” .................................... 55
D.
“providing a reference database of calculated locations of Wi-Fi access
points in a target area” ......................................................................................... 57
E.
“a WiFi-enabled device communicating with WiFi access points within
range of the WiFi-enabled device so that observed WiFi access points
identify themselves”............................................................................................. 57
F.
“in response to a user application request to determine a location of a userdevice having a Wi-Fi radio”............................................................................... 59
G.
“a user-device having a Wi-Fi radio” .................................................................. 60
H.
“simple signal strength weighted average model”............................................... 60
I.
“triangulation technique” ..................................................................................... 62

ii
A/74514379.11/3005005- 0000352152

I, Anthony S. Acampora, declare as follows:
1.

I am over 18 years of age. I have personal knowledge of the facts stated herein,

except those stated on information and belief, and, if called upon, could and would testify
competently to them. I make this declaration in support of Google Inc.’s Memorandum of Law
in Support of its Motion for Summary Judgment of Indefiniteness and, in the Alternative,
Opening Claim Construction Brief.
2.

I have been engaged as an expert to review issues relating to patents owned by

plaintiff Skyhook Wireless, Inc. (“Skyhook”), including the asserted patents, U.S. Patent Nos.
7,414,988 (“the ‘988 patent”), 7,433,694 (“ the ‘694 patent”), 7,474,897 (“the ‘897 patent), and
7,305,245 (“the ‘245 patent”) (collectively, “the Asserted Patents”) and for this declaration to
offer my opinion as to how a person of ordinary skill in the art would understand the Asserted
Patents.
3.

I discuss in this Declaration the meaning of certain language used in the claims of

the Asserted Patents. In forming my opinions as expressed below, I have relied primarily upon
the text of the claims and specifications of the Asserted Patents, their prosecution histories,
certain related patents and patent applications assigned to Skyhook, including their prosecution
histories, and exhibits to the Declaration of Susan Baker Manning In Support of Google Inc.’s
Motion for Summary Judgment of Indefiniteness and, in the Alternative, Opening Claim
Construction Brief, as well as my background, knowledge and experience in the field.
I.

QUALIFICATIONS AND ENGAGEMENT
4.

I have been asked to review the Skyhook patents and for purposes of claim

construction provide my opinions about what one of ordinary skill in the art would understand
terminology used in the claims of the patents to mean.
5.

In general, the patents relate to the collection of readings from wireless access

points and using those readings to map position information for the access points and using the

A/74514379.11/3005005-0000352152

readings and position information to determine the location of wireless devices. I have extensive
experience with wireless communications systems, including the location and characterization of
signals from wireless systems and use of such information to determine position.
6.

I received my Bachelor of Science, Master of Science, and Doctor of Philosophy

degrees, all in Electrical Engineering, from the Polytechnic Institute of Brooklyn in 1968, 1970,
and 1973, respectively. Both my Master’s thesis and my Ph.D. dissertation involved theoretical
aspects of electromagnetic wave propagation in plasma and gaseous media.
7.

From June 1968 through September 1988, I was employed at AT&T Bell

Laboratories in various engineering, research, and managerial positions, all in the general area of
telecommunications.
8.

My initial work at Bell Laboratories (1968-1974) involved high power radar

design and development, and signal design and processing for extraction of pertinent information
from radar target returns, both focused on anti-ballistic missile defense applications.
9.

My next assignment at Bell Laboratories (1974-1981) was in the Radio Research

Laboratory, an organization responsible for basic research, where I was involved in new
discovery and proposals involving novel approaches for communication satellite systems. My
contributions to the communication satellite state-of-the-art included (1) strategies to efficiently
encode and recover digital information sent to and from the satellites via high capacity radio
beams; (2) novel systems and on-board satellite switching approaches that use multiple radio
beams (so-called spot beams), each focused on a small portion of Earth, to vastly increase the
capacity of a communication satellite by enabling the radio spectrum to be re-used among the
spot beams; (3) strategies to acquire and maintain synchronization of radio signals sent to and
from a satellite; and (4) a novel approach to overcome the effects of rain-induced attenuation of

-2-

A/74514379.11/3005005-0000352152

the radio beams that dynamically adapts the modulation and assigns available radio resources to
those spots on Earth where rain attenuation is instantaneously most severe.
10.

I was promoted to Supervisor of the Data Theory Group at Bell Laboratories in

1981, with responsibility for exploratory development of local area data networks. These are
packet-switching networks intended to enable very high speed computer, voice, and video
communications via on-demand capture of a shared transmission channel.
11.

In 1984, I was promoted to Head of the Network Systems Research Department

(one of several departments within the Radio Research Laboratory, later to become the
Communications Systems Laboratory at Bell Laboratories) with responsibility for new
architectures for packet switching, multi-wavelength optical networks, wireless networks for
broadband local access, and integrated voice/data wireless networks. My contributions included
(1) a system architecture for using a raster of focused radio beams to deliver broadband service
to a large number of buildings from a central location within a city; (2) a novel packet switching
architecture for Internet-like wide area packet networks; and (3) a multi-wavelength multimedia
networking strategy to enable access to the enormous information-bearing capacity potential of
optical fiber cabling.
12.

I was promoted to Director of the Transmission Technology Laboratory in 1987, a

group of approximately 80 people with a broad charter for exploratory development of (1)
transmission and switching systems for next-generation Internet-like packet-based networks; and
(2) applications for digital signal processing in telecommunications.
13.

I left AT&T Bell Laboratories in September 1988 to become Professor of

Electrical Engineering and Director of the Center for Telecommunications Research at Columbia
University. Here, my responsibilities were three-fold: (1) education of students in the field of
telecommunications; (2) pursuit of a program of independent research in the area of
-3-

A/74514379.11/3005005-0000352152

telecommunications; and (3) management of a National Science Foundation Engineering
Research Center devoted to many aspects of telecommunications and founded for the express
purpose of improving American economic competitiveness through research, education, and
transfer of relevant technical findings from academia to the telecommunications industry.
Research programs at the Center for Telecommunications Research were focused on multiwavelength fiber optical networks, wireless communications, image and video communications,
network management and control, and underlying photonic and electronic devices and materials.
My contributions included (1) laboratory implementation and feasibility demonstration of the
world’s first multi-wavelength packet switched optical network; (2) new approaches for
randomly accessing a shared radio channel; (3) strategies for enabling rapid handoff among radio
cells in a high capacity cellular network; (4) a rigorous understanding of multi-wavelength
optical network capabilities and limitations; and (5) algorithms for the efficient resource
management and control of packet based multimedia networks.
14.

In August 1995, I left Columbia University to become Professor of Electrical and

Computer Engineering and Director of the Center for Wireless Communications at the
University of California, San Diego (UCSD). My responsibilities were three-fold: (1) education
of students in the field of wireless communications; (2) pursuit of a program of independent
research in the area of wireless communications; and (3) management of an industrially funded
research center devoted exclusively to wireless communications. My contributions included: (1)
strategies for allowing the use of so-called “smart” antennas in cellular-based packet radio
networks; (2) a proposal for a new city-wide network based on a wireless mesh-based approach
using either focused wireless beams of light or focused radio beams, intended to deliver
broadband services to buildings and/or to connect wireless radio cells with the worldwide fiberoptic backbone network; and (3) mobility management strategies for high speed packet-based
wireless networks. The second of these contributions has served as the technical foundation for
at least two new venture-backed telecommunications equipment companies, one of which I co-4-

A/74514379.11/3005005-0000352152

founded. In addition, I conducted further research on the subject of multi-wavelength optical
networks.
15.

In December 1999, I resigned as Director of the Center for Wireless

Communications to pursue full-time research and education as a Professor of Electrical and
Computer Engineering at UCSD, and on January 1, 2008, I became Professor of Electrical and
Computer Engineering, Emeritus, maintaining an active research program and teaching an
advanced graduate-level course on wireless networks.
16.

At UCSD, I have taught courses on (1) probability; (2) random processes; and (3)

wireless networks. My current research is focused on (1) broadband wireless networks for local
access to homes, schools, and businesses; (2) wireless spaces to enable ubiquitous voice, data,
and video wireless communications within buildings; and (3) so-called ad-hoc (self organizing)
networks of wireless sensor nodes for business and homeland security applications.
17.

Over the course of my career, I have published (individually or with

collaborators) over 170 original papers in scholarly journals and professional conference
proceedings. I am the named inventor or co-inventor on 39 U.S. patents. I wrote one of the
world’s first textbooks devoted to broadband telecommunications, titled An Introduction to
Broadband Networks.
18.

I have lectured extensively on telecommunications, in general, and wireless

communications in particular. I have regularly attended, and continue to attend, numerous
worldwide professional conferences. I have chaired several major telecommunications
conferences, and I have chaired numerous professional conference technical sessions. I read the
technical literature extensively and subscribe to several leading journals in the field of
telecommunications, in general, and wireless communications in particular. Over the years, I
have delivered many three-to-five-day intensive short courses on telecommunications and
-5-

A/74514379.11/3005005-0000352152

wireless communications to professional audiences of practicing engineers and others. In 1988, I
was elected to the grade of Fellow of the Institute of Electrical and Electronics Engineers, cited
for contributions to high-capacity digital satellite systems and broadband local communication
networks.
19.

My curriculum vitae, which summarizes my professional background, experience

and publications, is attached as Exhibit 1 hereto.
II.

THE PATENTS-IN-SUIT
A.

Background and Description of the Claimed Invention.

20.

The patents-in-suit generally address the creation and application of a database

tied to a geographic area that might be used to determine the location of the user of a Wi-Fi
device. The inventors rely on the creation and updating of a database, by systematically
scanning for access points, that contains the estimated locations of all Wi-Fi access points within
some geographically dispersed target region. Then, by responding to user queries reporting the
identities of those access stations that are observable to the user of a Wi-Fi device, the database
provides location information for the access points used to estimate the location of the Wi-Fi
device itself.
21.

It appears that the inventors were unhappy with the existing systems and methods

for accomplishing these same two objectives: creation of, and application of, a database of WiFi access points within some target region. Their “discovery,” if any, appears to be a deliberate,
and possibly unachievable, effort to improve the accuracy with which the location of a Wi-Fi
device might be determined by improving the collection process to only scan Wi-Fi access points
as described in the patents. What they apparently believe that they “discovered” is what they
describe as a more accurate way to create a database of Wi-Fi access point locations by
deliberately taking measurements in some sort of systematic fashion. It would appear from the
disclosure that they believe that they accomplished this by intentionally only scanning for access
-6-

A/74514379.11/3005005-0000352152

points along every street within the target region using an algorithm to determine the most
efficient route along every street.
22.

The specification is dominated by a single method for creating the database,

which involves driving a vehicle in a systematic manner along every street, and scanning for
“beacons” broadcast from surrounding access points. The scanning is performed using a Chinese
Postman format to drive each street a minimum number of times, and preferably only once to
avoid introducing a bias toward certain streets. Each beacon identifies the access point from
which it was broadcast. The signal strength of the observed beacon is also recorded, as is the
precise location of the vehicle at the time that the beacon observation and measurement was
made. This precise location is obtained from Global Positioning System (GPS) equipment
installed in the vehicle.
23.

The beacon sent by a particular access point may be observed several times as the

van travels along every street in the target region. For each beacon observation and signal
strength measurement, the location of the access point is assumed to be the GPS location of the
vehicle. Then, using a process that the inventors refer to as “reverse triangulation,” an estimated
location of the access point is computed by doing an average of the recorded GPS locations, with
each GPS location weighted in a nonlinear fashion by the signal strength of the accompanying
measurement. The only description of this “reverse triangulation” appears to be the set of
equations appearing in column 12 of the ‘694 patent. These equations also appear in the ‘245
and ‘988 patents. Absent these equations, one of skill in the art would not know what is meant
by “reverse triangulation” as the term is used in the patents. The desirability of deleting
anomalous access point location measurements before reverse triangulation is discussed.
24.

To determine the location of a Wi-Fi device, the device first listens to beacons

sent by surrounding access points. The device then queries the database, which provides the
estimated locations of the corresponding access points. These estimated access point locations,
-7-

A/74514379.11/3005005-0000352152

along with the observed signal strengths of the measured beacons, are used in another application
of “reverse triangulation” to produce a computed location of the Wi-Fi device. Once again, the
only description of this process appearing in the patents are the set of equations appearing in
column 12 of the ‘694 patent; as previously noted, these equations also appear in the ‘245 and
‘988 patents.
B.

The asserted claims.

25.

It is my understanding that Skyhook accuses Google of infringing claims 1-3 of

the ‘988 patent; claims 1 and 2 of the ‘694 patent; claims 1-4 of the ‘897 patent; and claims 1, 2,
4-6, and 8 of the ‘245 patent. Those claims recite:
The ‘988 patent
1.

A Wi-Fi location server, comprising:

a database of Wi-Fi access points for at least one target area having
a radius on the order of tens of miles,
said database being recorded in a computer-readable medium and
including database records for substantially all Wi-Fi access points
in the target area,
each record including identification information for a
corresponding Wi-Fi access point and calculated position
information for the corresponding Wi-Fi access point,
wherein said calculated position information is obtained from
recording multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point so that the multiple
readings have reference symmetry relative to other Wi-Fi access
points in the target area and so that the calculation of the position
of the Wi-Fi access point avoids arterial bias in the calculated
position information; and
computer-implemented logic to add records to the database for
newly-discovered Wi-Fi access points said computer logic
including logic to recalculate position information for Wi-Fi access
points previously stored in the database to utilize position
information for the newly-discovered readings of previously stored
Wi-Fi access points.
2.
The server of claim 1 further including computerimplemented clustering logic to identify position information
based on error prone GPS information.
-8-

A/74514379.11/3005005-0000352152

3.
The server of claim 2 wherein the clustering logic includes
logic to determine a weighted centroid position for all position
information reported for an access point and logic to identify
position information that exceeds a statistically-45 based deviation
threshold amount away from the centroid position and excludes
such deviating position information from the database and from
influencing the calculated positions of the Wi-Fi access points.
The ‘694 patent
1.
A database of Wi-Fi access points for at least one target
area having a radius on the order of tens of miles,
said database being recorded in a computer-readable medium and
including database records for substantially all Wi-Fi access points
in the target area,
each record including identification information for a
corresponding Wi-Fi access point and calculated position
information for the corresponding Wi-Fi access point,
wherein said calculated position information is obtained from
recording multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point so that the multiple
readings avoid arterial bias in the calculated position information
of the Wi-Fi access point, and
wherein the database records for substantially all Wi-Fi access
points in the target area provide reference symmetry within the
target area.
2.
The database of claim 1 having database records for a
plurality of target areas, said database records being organized by
target areas.
The ‘897 patent
1.
In a location-based services system for WiFi-enabled
devices, a method of calculating the position of WiFi-enabled
devices comprising the acts of:
a)
a WiFi-enabled device communicating with WiFi access
points within range of the WiFi-enabled device so that observed
WiFi access points identify themselves;
b)
accessing a reference database to obtain information
specifying a recorded location for each observed WiFi access
point;
c)
using the recorded location information for each of the
observed WiFi access points in conjunction with predefined rules
-9-

A/74514379.11/3005005-0000352152

to determine whether an observed WiFi access point should be
included or excluded from a set of WiFi access points;
d)
using the recorded location information of only the WiFi
access points included in the set and omitting the recorded location
information of the excluded WiFi access points to calculate the
geographical position of the WiFi-enabled device.
2.
The method of claim 1 further including recording signal
strength information for WiFi access points included in the set and
using the signal strength information when calculating the
geographical position of the WiFi-enabled device.
3.
The method of claim 1 wherein the predefined rules include
rules to determine a reference point and to compare the recorded
location information for each of the observed WiFi access points to
the reference point, and wherein WiFi access points having a
recorded location within a predefined threshold distance of the
reference point are included in the set and wherein WiFi access
points having a recorded location in excess of the predefined
threshold distance of the reference point are excluded from the set.
4.
The method of claim 3 wherein the reference point is
determined by identifying a cluster of WiFi access points and
determining an average position of the WiFi access points in the
cluster.
The ‘245 patent
1.
A method of locating a user-device having a Wi-Fi radio,
comprising:
providing a reference database of calculated locations of Wi-Fi
access points in a target area;
in response to a user application request to determine a location of
a user-device having a Wi-Fi radio, triggering the Wi-Fi device to
transmit a request to all Wi-Fi access points within range of the
Wi-Fi device;
receiving messages from the Wi-Fi access points within range of
the Wi-Fi device, each message identifying the Wi-Fi access point
sending the message;
calculating the signal strength of the messages received by the WiFi access points;
accessing the reference database to obtain the calculated locations
for the identified Wi-Fi access points;
based on the number of Wi-Fi access points identified via received
messages, choosing a corresponding location determination
- 10 -

A/74514379.11/3005005-0000352152

algorithm from a plurality of location determination algorithms,
said chosen algorithm being suited for the number of identified
Wi-Fi access points;
using the calculated locations for the identified Wi-Fi access points
and the signal strengths of said received messages and the chosen
location-determination algorithm to determine the location of the
user-device.
2.
The method of claim 1 wherein the calculated locations for
the identified Wi-Fi access points are filtered to determine if the
corresponding Wi-Fi access points have moved since the time the
information about the Wi-Fi access points was included in the
reference database.
4.
The method of claim 1 wherein the reference database is
located remotely relative to the user-device.
5.
The method of claim 1 wherein the location of the user
device is provided with latitude and longitude coordinates.
6.
The method of claim 1 wherein the plurality of locationdetermination algorithms includes a simple signal strength
weighted average model.
8.
The method of claim 1 wherein the plurality of locationdetermination algorithms includes a triangulation technique.
C.

The asserted patents’ familial relationship to each other.

26.

The Asserted Patents are all closely related and three claim priority to a common

provisional patent as discussed below. Each identifies the same four individuals as inventors
(Russel Kipp Jones, Farshid Alizadeh-Shabdiz, Edward James Morgan, and Michael George
Shean), and each states on its face that it is assigned to plaintiff Skyhook Wireless, Inc.
27.

On their face, the ‘988, ‘694 and ‘245 patents each claims priority to U.S.

Provisional Application No. 60/623,108, which was filed with the U.S. Patent and Trademark
Office on October 29, 2004. The applications that later issued as the ‘988, ‘694 and ‘245 patents
were filed on October 28, 2005.
28.

In addition to each claiming priority to the same provisional application, the ‘988,

‘694 and ‘245 patents each state that they are related to the others, as well as to the unasserted

- 11 -

A/74514379.11/3005005-0000352152

‘762 patent. See ‘988 patent at Col. 1, lines 12-22; ‘694 patent at Col. 1, lines 11-32; and ‘245
patent at Col. 1, lines 14-19.
29.

The ‘897 patent is also part of the same patent family. It issued from a February

22, 2006 application that claims priority as a continuation-in-part of the application that issued as
the ‘245 patent. The ‘897 patent includes similar language identifying the relationship to the
other family members: U.S. Provisional Application No. 60/654,811 (filed on February 22,
2005); U.S. Provisional Application No. 60/658,481 (filed on Mar. 4, 2005); the application that
issued as the ‘988 patent (asserted); the application that issued as the ‘694 patent (asserted); the
application that issued as the ‘245 patent (asserted); the application that issued as U.S. Patent No.
7,403,762 (unasserted); the application that issued on February 19, 2009 as the U.S. Patent No.
7,493,127 (unasserted); and pending U.S. Patent App. No. 11/359,154 (filed Feb. 22, 2006).
D.

The alleged invention of the asserted claims—”Arterial Bias,” “Reference
Symmetry,” and “Chinese Postman.”

30.

Generally speaking, the ‘988 and ‘694 patents both claim a database of calculated

position information for Wi-Fi access points in a target area. As recited in the claims, and as
taught in the specification, that calculated position information is derived from raw scanning data
that was collected using a certain driving methodology. Although the claims and specifications
of the ‘988 and ‘694 patents are profoundly unclear in important respects, as I discuss in greater
detail below, the general concept appears to be that the “Chinese Postman” routing methodology
taught and claimed addresses the problem of “arterial bias” described in the specifications, and is
related to the provision of “reference symmetry” as also taught and claimed.
31.

Claim 1 of the ‘988 is similar to claim 1 of the ‘694 patent. As shown in the side-

by-side comparison below, the first three limitations of each are identical. The fourth limitation
of claim 1 of the ‘988 shares much of the same language as the fourth and fifth limitations of
claim 1 of the ‘694 patent, although they are not identical in every respect. The fifth limitation

- 12 -

A/74514379.11/3005005-0000352152

of claim 1 of the ‘988 recites requirements for claimed “computer-implemented logic” that is not
paralleled in claim 1 of the ‘694 patent.
988/1
A Wi-Fi location server, comprising:

694/1

a database of Wi-Fi access points for at least
one target area having a radius on the order of
tens of miles,
said database being recorded in a computerreadable medium and including database
records for substantially all Wi-Fi access
points in the target area,
each record including identification
information for a corresponding Wi-Fi access
point and calculated position information for
the corresponding Wi-Fi access point,
wherein said calculated position information is
obtained from recording multiple readings of
the Wi-Fi access point at different locations
around the Wi-Fi access point so that the
multiple readings have reference symmetry
relative to other Wi-Fi access points in the
target area and so that the calculation of the
position of the Wi-Fi access point avoids
arterial bias in the calculated position
information; and
computer-implemented logic to add records to
the database for newly-discovered Wi-Fi
access points said computer logic including
logic to recalculate position information for
Wi-Fi access points previously stored in the
database to utilize position information for the
newly-discovered readings of previously stored
Wi-Fi access points.
32.

A database of Wi-Fi access points for at least
one target area having a radius on the order of
tens of miles,
said database being recorded in a computerreadable medium and including database
records for substantially all Wi-Fi access
points in the target area,
each record including identification
information for a corresponding Wi-Fi access
point and calculated position information for
the corresponding Wi-Fi access point,
wherein said calculated position information is
obtained from recording multiple readings of
the Wi-Fi access point at different locations
around the Wi-Fi access point so that the
multiple readings avoid arterial bias in the
calculated position information of the Wi-Fi
access point, and

wherein the database records for substantially
all Wi-Fi access points in the target area
provide reference symmetry within the target
area.

The specification of the ‘988 patent is very similar to that of the ‘694 patent. See

Ex. K .1 The two patents share the same eleven figures. The two detailed descriptions of the
inventions are identical, using exactly the same language to describe collection of Wi-Fi access
point data using the “Chinese Postman” routing methodology to obtain reference symmetry

1

Unless otherwise indicated, all cited exhibits are attached to the Declaration of Susan Baker Manning.

- 13 -

A/74514379.11/3005005-0000352152

while avoiding arterial bias. Ex. K at 6-10; Ex. C at Fig. 2, Fig. 11, 8:28-34, 8:44-59, 9:64-10:4.
They also describe in the same way processing and quality checking the data to be added to the
database of Wi-Fi access points. Ex. K at 12-14 & Ex. C at 11:46-14:7.
33.

The ‘245 patent relates to a method of locating a user-device by using a reference

database of Wi-Fi access points. It also discloses a method of building the reference database by
collecting Wi-Fi access point information while driving a “Chinese Postman” route to avoid
arterial bias and maintain reference symmetry.
34.

The specification of the ‘245 patent is likewise very similar to that of the ‘988 and

‘694 patents, though they are not identical in that the summaries of the inventions and
discussions of related art differ. In other respects they are the same, sharing the same figures and
detailed descriptions, including details regarding collection of Wi-Fi access point data using the
“Chinese Postman” routing methodology to try to differentiate collection methods acknowledged
in the prior art. Ex. L.
35.

The ‘897 patent contains additional disclosures beyond the ‘245 patent from

which it claims priority, and both the ‘897 patent and the ‘245 patent relate to a method of
calculating the position of a Wi-Fi enabled user device using a reference database. The ‘897
patent reiterates that a fault with prior art systems is the arterial bias introduced when data is
collected by individuals who do not follow designed scanning routes. The claims of both patents
disclose methods of calculating position information. The patents claim slightly different aspects
of the process of determining location of a Wi-Fi enabled device; the ‘897 patent claims predefined rules for including and excluding observed access points from a set used to determine
location, while the ‘245 patent claims a method of choosing amongst algorithms for location
determination. The specification of the ‘245 patent discloses the use of the same “Chinese
Postman” routing methodology for collection of access point data disclosed in the ‘694 and ‘988

- 14 -

A/74514379.11/3005005-0000352152

patents, while the ‘897 lists arterial bias and lack of reference symmetry among reference points
as drawbacks in the related art.
E.

The Prosecution History of the ‘988 Patent

36.

On November 30, 2007, the Examiner rejected pending claim 1 in the application

for the ‘988 patent as obvious in light of U.S. Patent App. Pub. No. 2005/0164710 (Beuck) in
view of U.S. Patent App. Pub. No. 2005/0037775 (Moeglein). Ex. G at GSHFED200-12. The
Examiner also objected to claim 1 because the term “radius on the order of tens of miles” “leaves
the claim open ended.” Id. at GSHFED202. The Examiner also rejected pending claims 2 and 3
as unpatentable in light of the Beuck reference in view of Moeglein and U.S. Patent No.
5,940,825 (Castelli). Id. at GSHFED207-10.
37.

In response, the applicants amended the last two limitations of claim 1:
A Wi-Fi location server, comprising:
A database of Wi-Fi access points for at least one target area
having a radius on the order of tens of miles,
said database being recorded in a computer-readable medium and
including database records for substantially all Wi-Fi access points
in the target area,
each record including identification information for a
corresponding Wi-Fi access point and calculated position
information for the corresponding Wi-Fi access point,
wherein said calculated position information is obtained from
recording multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point so that the multiple
readings have to provide reference symmetry relative to other WiFi access points in the target area when calculating and so that the
calculation of the position of the Wi-Fi access point and to avoids
arterial bias in the calculated position information; and
computer-implemented logic to add records to the database for
newly-discovered Wi-Fi access points said computer logic
including logic to recalculate position information for Wi-Fi access
points previously stored in the database to utilize position
information for the newly-discovered readings of previously stored
Wi-Fi access points.
- 15 -

A/74514379.11/3005005-0000352152

Ex. G at GSHFED183.
38.

The amendments are all directed to obtaining the claimed “calculated position

information.” As amended, the limitation both specifies that the “calculated position
information” has certain characteristics, and specifies how the data used to calculate the location
of Wi-Fi access points is collected. According to the claim language, the “calculated position
information” must “avoid arterial bias.” To obtain the data used to calculate the location of the
Wi-Fi access points one must “record[] multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point.” According to the claim, this collection method results
in the “multiple readings hav[ing] reference symmetry relative to other Wi-Fi access points in the
target area.”
39.

The applicants also provided detailed remarks in which they argued that the

amended claims were patentable over the prior art. Ex. G at GSHFED185-91. In particular, the
applicants argued that the prior art, unlike the invention, was silent as to the methodology of
determining the location of the Wi-Fi access points.
In contrast to the cited references, applicants’ claim 1 is directed to
a Wi-Fi location server that includes position information for WiFi access points without arterial bias. Specifically, the calculated
position information for the Wi-Fi access points is obtained from
recording multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point. These multiple readings
have reference symmetry relative to other Wi-Fi access points in
the target area. Thus, the calculation of the position of the Wi-Fi
access point avoids arterial bias in the calculated position
information. This technique of gathering readings from Wi-Fi
access points results in higher quality estimates of access point
locations and more complete information about the access points in
the area. Consequently, devices using the calculated access point
locations to determine their position have more accurate
estimations of their locations. See Application at ¶¶ 41-44.
As set forth above, none of the cited reference teach or suggest
conducting an audit of an area to build a reference database of
the locations of Wi-Fi access points in a target area so as to
provide reference symmetry and avoid arterial bias. As stated
in the application, amateur scanners (“wardrivers”) have attempted
to collect access point location data for use in location estimation
- 16 -

A/74514379.11/3005005-0000352152

systems. However, the methods employed by wardrivers suffer
from several drawbacks. Namely, as described in the
application, the location data collected by the wardrivers is
often inaccurate, incomplete, and grows organically rather
than being collected in a systematic fashion to purposefully
avoid arterial bias. See Application at ¶¶ 15-17.
As explained in greater detail in the application, significant errors
in position calculation can result when the reference points used
for the calculation lack symmetry around the physical location of
the device performing the calculation. Unsymmetrical location
data (or “arterial bias”) occurs when individuals (e.g.,
wardrivers) collect location data for Wi-Fi access points
without following designated scanning routes. Such data tends
to aggregate around heavily traffic areas (or “arteries”).
Attempting to use arterially biased data to estimate the location of
a mobile device causes a “location pull” towards the main arteries
regardless of where the user is currently located. This causes
substantial accuracy errors in the location estimation. Figures 5 and
6 of the application illustrate this effect. See Application at ¶¶ 15
and 44.
Collecting multiple readings of Wi-Fi access points in a
systematic fashion, as described in the application, provides
reference symmetry within the target area. Thus, the
distribution of reference points (i.e., Wi-Fi access point locations)
is symmetric. By using a collection of location data that is
symmetric, a mobile device attempting to calculate its location
typically encounters physical locations in which there are
numerous access point locations on all sides of the device within
range of the device’s Wi-Fi radio. Therefore, a position calculation
performed by the mobile device will have reduced location bias
and will be more accurate as a result. See Application at ¶ 44.
Unlike the cited references and known methods described in the
background of the application, applicants’ claim 1 clearly recites
the calculated position information is obtained from recording
multiple readings of the Wi-Fi access point at different locations
around the Wi-Fi access point so that the multiple readings have
reference symmetry relative to other Wi-Fi access points in the
target area and so that the calculation of the position of the Wi-Fi
access point avoids arterial bias in the calculated position
information. The application describes the discovery of the arterial
bias problem and the advantages of the solutions devised by
applicants. Namely, by performing a planned audit, and
avoiding arterial bias, applicants at least achieve more
complete information about access points in the target area,
higher quality estimates of access point locations, and
reference symmetry. See Application at ¶¶ 47-51.
None of this is taught or suggested by the cited references. Thus,
applicants submit that claim 1 is patentable over the cited
references.
- 17 -

A/74514379.11/3005005-0000352152

Ex. G at GSHFED0000187-89 (emphasis added).
40.

In light of these remarks—and also consistent with the specification and the

amendments—it would be apparent to a person of ordinary skill that Skyhook’s claims were not
intended to cover unsystematic methods of collecting information about Wi-Fi access points.
41.

Consistent with that teaching, the applicants emphasized to the Examiner that the

invention solved the problem of arterial bias, and ensured reference symmetry, due to the method
by which the data is collected. The data cannot be collected in a random manner.
•

The applicants twice referred to an “audit” or “planned audit,” which itself clearly
conveys that the target area is to be traversed systematically. See Ex. G at
GSHFED0000187; id. at GSHFED0000189.

•

The applicants also twice referred to collecting the location information in a
“systematic fashion.” See Ex. G at GSHFED0000188.

•

Scanning is done by a “fleet of vehicles [that] perform their scanning routines
while driving their pre-designed routes.” Ex. C at 11:49-50.

•

The applicants criticized data collection “without following designated scanning
routes” as leading to arterial bias. See Ex. G at GSHFED0000188.

42.

The specification also criticizes the “Random Scanning Model” at length. See,

e.g., Ex. C at 2:44-3:27, Figs. 3 &11. It describes the “Random Scanning Model” as leading to
the problem of inaccurate location calculations due to arterial bias; systematic location data
collection by following a Chinese Postman route is the applicants’ solution to that problem. See
id. 8:28-59, 9:6-21, 9:57-10:4.
43.

The claims and prosecution history each teach that avoidance of arterial bias must

be a specific goal of the data collection method:
•

“wherein said calculated position information is obtained from recording multiple
readings of the Wi-Fi access point at different locations around the Wi-Fi access
point so that the multiple readings have reference symmetry relative to other WiFi access points in the target area and so that the calculation of the position of the
- 18 -

A/74514379.11/3005005-0000352152

Wi-Fi access point avoids arterial bias in the calculated position information”
(‘988 patent, claim 1);
•

“[N]one of the cited reference teach or suggest conducting an audit of an area to
build a reference database of the locations of Wi-Fi access points in a target area
so as to provide reference symmetry and avoid arterial bias.” Ex. G at
GSHFED0000187 (emphasis added); and

•

“[A]s described in the application, the location data collected by the wardrivers is
often inaccurate, incomplete, and grows organically rather than being collected in
a systematic fashion to purposefully avoid arterial bias. See Application at ¶¶ 1517.2“ Ex. G at GSHFED0000188 (emphasis added).

44.

The Examiner allowed the claims on May 5, 2008. In doing so, he did not give

any detailed reasoning, merely stating that amended claim 1, which he quoted verbatim, was
patentable over two prior art references. He did not comment on his earlier rejection of claim 1
as “open ended.”
45.

The term “calculated position information” and similar terminology used in the

patents is used to refer to data that is collected for access points and used to determine a location
for each access point. One of ordinary skill in the art would understand from reading the ‘988
patent and related patents that:
a. “calculated position information” must be based on raw scanning data (but is not
the raw scanning data itself, per “calculated”)
b. “calculated position information” is about the physical location (lat & long) of the
Wi-Fi access point
c. The raw scanning data must have been collected during an audit of the target area
in which drivers followed a Chinese Postman route.

2

Paragraphs 15-17 of the Application for the ‘988 patent correspond to unamended claims 1-3, which are
recited in the Summary at column 4, lines 28-57. The applicants thus directly tied the concepts of
systematic data collection and purposeful avoidance of arterial bias to the claims.

- 19 -

A/74514379.11/3005005-0000352152

d. The collection of the raw scanning data must have been intended to avoid arterial
bias in the calculated location.
e. The raw scanning data cannot have been collected by end-users.
f. The raw scanning data cannot have been collected by the random scanning method
disparaged in the specification, and distinguished during prosecution.
These principles apply equally to ‘694 patent (in which the term “calculated position
information” also appears), as well as to the construction of “calculated locations” in the related
‘245 patent and “recorded location information” in the related ‘897 patent.
46.

As to the Examiner’s objection to the “radius on the order of tens of miles”

limitation as “leav[ing] the claim open ended,” the Applicants argued that the limitation “clearly
communicates that the claimed target area is larger than, for example, a single floor of a
building, such as might be found in an indoor positioning system. See Application at ¶ 16.
Applicants describe throughout the application an embodiment that includes position information
for Wi-Fi access points within a large metropolitan area.” Ex. G at GSHFED190.
47.

The Examiner issued a Notice of Allowability on May 21, 2008. Ex. G at

GSHFED168-72. The ‘988 patent issued on August 19, 2008.
F.

The Prosecution History of the ‘694 Patent

48.

During the prosecution of the ‘694 patent, the Examiner rejected claims 1 and 2

under 35 U.S.C. § 103(a) as being unpatentable over U.S. Patent Application Publication No.
2004/0039520 (Khavakh) in view of U.S. Patent Application Publication No. 2004/0058640
(Root). Ex. H at GSHFED311. Khavakh teaches a database of Wi-Fi access points recorded on a
computer-readable medium, each record containing calculated position information for each WiFi access point, and calculated position information obtained from multiple readings of Wi-Fi
access points to provide reference symmetry and to avoid arterial bias. Id. Root teaches having
a radius on the order of tens of miles. Id. The examiner determined that it would have been
obvious to provide the teaching of Root into the system of Khavakh to predict events within a
- 20 -

A/74514379.11/3005005-0000352152

particular special range of a particular dynamic special location; therefore claim 1 was rejected.
Ex. H at GSHFED312. Claim 2 was rejected because the combination of Khavakh and Root
teaches the database of claim 1 having records for a plurality of target areas, organized by target
areas. Id.
49.

On April 7, 2008, the Applicants apparently held a telephonic interview with the

Examiner. Ex. H at GSHFED298.
50.

One day later, on April 8, 2008, the Applicants amended claim 1 to “more

particularly recite characteristics of the calculated position information,” and submitted that the
amendments overcome the rejection. Ex. H at GSHFED295-99. Specifically, the Applicants
amended the fourth limitation of claim 1 of the ‘694 patent (regarding the avoidance of arterial
bias) and added the fifth limitation (regarding the provision of reference symmetry):
A database of Wi-Fi access points for at least one target area
having a radius on the order of tens of miles,
said database being recorded in a computer-readable medium and
including database records for substantially all Wi-Fi access points
in the target area,
each record including identification information for a
corresponding Wi-Fi access point and calculated position
information for the corresponding Wi-Fi access point,
wherein said calculated position information is obtained from
recording multiple readings of the Wi-Fi access point at different
locations around the Wi-Fi access point so that the multiple
readings to provide reference symmetry when calculating the
position of the Wi-Fi access point and to avoid arterial bias in the
calculated position information of the Wi-Fi access point, and
wherein the database records for substantially all Wi-Fi access
points in the target area provide reference symmetry within the
target area.
Ex. H at GSHFED297.

- 21 -

A/74514379.11/3005005-0000352152

51.

According to the Applicants’ Remarks accompanying the Amendment, “During

the telephone call, applicants submitted that the cited references do not teach or suggest these
features [i.e., the claims as amended]. Examiner Danh stated that the amendments overcome the
cited references.” Ex. H at GSHFED298.
52.

The Examiner issued a Notice of Allowability on June 16, 2008. Ex. H at

GSHFED282-284. The ‘694 patent issued on October 7, 2008.
G.

The Prosecution Histories of the ‘245 and ‘897 Patents

53.

The Examiner allowed both the ‘245 and ‘897 patents to issue with the original

claims as-filed. Ex. I at GSHFED87-90 (September 12, 2007 Notice of Allowabilty re ‘245
patent); Ex. J at GSHFED392-95 (August 14, 2008 Notice of Allowabilty re ‘897 patent).
54.

In allowing the ‘245 patent, Examiner Le identified the limitation “based on the

number of Wi-Fi access points identified via received messages, choosing a corresponding
location-determination algorithm from a plurality of location-determination algorithms, said
chosen algorithm being suited for the number of identified Wi-Fi access points” as the point of
novelty over seven identified U.S. patents or published applications (Masouka, Krumm,
Meunier, Patil, Sheynblat, Vesuna, and Reeves). See Ex. I at GSHFED0000089-90.
55.

In allowing the ‘897 patent eleven months later, Examiner Le identified steps

c) and d) of claim 1 as the point of novelty over the prior art. See Ex. J at GSHFED0000394-95
(noting the Choti, Agrawa, Orwant, Biffar, Nagda, and Zellner references). Those limitations
recite:
c) using the recorded location information for each of the observed
WiFi access points in conjunction with predefined rules to
determine whether an observed WiFi access point should be
included or excluded from a set of WiFi access points
d) using the recorded location information of only the WiFi access
points included in the set and omitting the recorded location
information of the excluded WiFi access points to calculate the
geographical position of the WiFi-enabled device
- 22 -

A/74514379.11/3005005-0000352152

‘897 patent, claim 1.
III.

SUMMARY OF STIPULATED AND DISPUTED CLAIM CONSTRUCTIONS.
56.

I am advised that Google and Skyhook agree that the following constructions

should be adopted in this case:
Claim language

“Wi-Fi access points”

Stipulated construction:

Devices operating consistent with the
IEEE 802.11 standard to provide
network connectivity.

“a radius on the order of tens of miles” A radius of ten miles or more but
fewer than a hundred miles.
“identification information for a
corresponding Wi-Fi access point”

An identifier (e.g., a MAC address) for
a corresponding Wi-Fi access point.

Appears in
Asserted
Claims
‘988/1, 3
‘694/1
‘245/1, 2
‘897/1, 3, 4
‘988/1
‘694/1
‘988/1
‘694/1

See “Wi-Fi access points.”

“triggering the Wi-Fi device to
transmit a request to all Wi-Fi access
points within range of the Wi-Fi
device”

Causing the Wi-Fi device to actively
search for Wi-Fi access points. The WiFi device transmits a request to all WiFi access points within range of the WiFi device to identify themselves.

‘245/1

See “Wi-Fi access points.”
“WiFi access points having a recorded
location within a predefined threshold
distance of the reference point”

WiFi access points having a recorded
‘897/3
location that is within a certain distance
of the reference point. That distance was
previously defined.
See “WiFi access points.”

“WiFi access points having a recorded
location in excess of the predefined
threshold distance of the reference
point”

WiFi access points having a recorded
location that exceeds a certain distance
from the reference point. That distance
was previously defined.

‘897/3

See “WiFi access points.”

57.

I am advised that Google and Skyhook do not agree as to the proper construction

of the following terms:

- 23 -

A/74514379.11/3005005-0000352152

Claim language
“target area”

Appears in
asserted
claims
‘988/1
‘694/1, 2
‘245/1

“substantially all Wi-Fi access points” / “for substantially all Wi-Fi access
points in the target area”

‘988/1
‘694/1

“calculated position information”

‘988/1
‘694/1

“reference symmetry”
“arterial bias”

‘988/1
‘694/1

“recording multiple readings of the Wi-Fi access point at different locations
around the Wi-Fi access point so that the multiple readings have reference
symmetry relative to other Wi-Fi access points in the target area”

‘988/1

“recording multiple readings of the Wi-Fi access point at different locations
around the Wi-Fi access point … so that the calculation of the position of the
Wi-Fi access point avoids arterial bias in the calculated position information”

‘988/1

“avoid arterial bias” /”avoids arterial bias”

‘988/1
‘694/1

“logic to recalculate position information for Wi-Fi access points previously
stored in the database to utilize position information for the newly-discovered
readings of previously stored Wi-Fi access points”

‘988/1

“computer-implemented logic to add records to the database for newlydiscovered Wi-Fi access points”

‘988/1

“computer-implemented clustering logic to identify position information based
on error prone GPS information”

‘988/2

“logic to determine a weighted centroid position for all position information
reported for an access point”

‘988/3

“a weighted centroid position”

‘988/3

“logic to identify position information that exceeds a statistically-based
deviation threshold amount away from the centroid position”

‘988/3

“calculated positions of the Wi-Fi access points”

‘988/3

- 24 -

A/74514379.11/3005005-0000352152

Claim language
“the clustering logic … excludes such deviating position information from the
database and from influencing the calculated positions of the Wi-Fi access
points”

Appears in
asserted
claims
‘988/3

“recording multiple readings of the Wi-Fi access point at different locations
‘694/1
around the Wi-Fi access point so that the multiple readings avoid arterial bias in
the calculated position information of the Wi-Fi access point”
“wherein the database records for substantially all Wi-Fi access points in the
target area provide reference symmetry within the target area”

‘694/1

“a user-device having a Wi-Fi radio”

‘245/1

“providing a reference database of calculated locations of Wi-Fi access points
in a target area”

‘245/1

“calculated locations”

‘245/1, 2

“in response to a user application request to determine a location of a userdevice having a Wi-Fi radio”

‘245/1

“said chosen algorithm being suited for the number of identified Wi-Fi access
points”

‘245/1

“simple signal strength weighted average model”

‘245/6

“triangulation technique”

‘245/8

“a WiFi-enabled device communicating with WiFi access points within range
of the WiFi-enabled device so that observed WiFi access points identify
themselves”

‘897/1

“using the recorded location information for each of the observed WiFi access
points in conjunction with predefined rules to determine whether an observed
WiFi access point should be included or excluded from a set of WiFi access
points”

‘897/1

“recorded location information”

‘897/1, 3

“rules to determine a reference point and to compare the recorded location
information for each of the observed WiFi access points to the reference point”

‘897/3

- 25 -

A/74514379.11/3005005-0000352152

IV.

CLAIM CONSTRUCTION PRINCIPLES
58.

It is my understanding that the claim construction process is governed by a

number of legal principles. Although I am not a lawyer, I have been advised on the relevant
legal principles, and have used them in considering whether the disputed limitations in
Skyhook’s patents reasonably can be given any definition, and if so what the appropriate
definition would be.
A.

The Level of Ordinary Skill in the Art

59.

It is my understanding that words in the claims of a patent are to be interpreted

according to their meaning to one of ordinary skill in the art, unless the terms have been given a
special meaning in the patent or in related documents such as the prosecution history.
60.

In this case, it is my opinion that one of ordinary skill in the art would be one with

a bachelor’s degree in electrical engineering or computer science and 3 – 5 years of experience
working in wireless communications hardware and software design.
61.

In forming my opinion regarding the education, skill level and background of a

person of ordinary skill in the relevant art, I considered a number of factors including the field of
the purported invention of the patents-in-suit, the skill required for implementing wireless
communications hardware and software, and my experience in wireless communications. This
opinion, like all of my opinions expressed in this report, is also based on my background,
education, and experience in the field of wireless communications.
B.

Specification, Prosecution History

62.

It is my understanding that the person of skill in the art is presumed to have read

and understood the claim terms and the specification. Although the claims, and not the
specification, define the scope of the patented invention, the specification is highly relevant to
the proper understanding of the claims. In fact, I understand that the courts have declared that
the specification is the single best guide to the meaning of a disputed term. Properly understood,
the ordinary meaning of the claim terms is therefore not their ordinary meaning in some abstract
- 26 -

A/74514379.11/3005005-0000352152

sense, but rather their ordinary meaning in the context of the entire patent, including the text of
the specification, the figures and the other claims. As discussed below, I have therefore carefully
considered the specifications of Skyhook’s patents, and what they indicate about the inventions
claimed in the various patents.
63.

I understand that any proper analysis of the meaning of the claims also requires

one to consider the prosecution history of the patent. I am also informed that it is appropriate to
consider the prosecution histories of patents that are related to the patent being considered to the
extent that such related prosecution histories shed light on the particular terms at issue. It is my
understanding that statements made by a patent applicant to the U.S. Patent and Trademark
Office can bear directly on the meaning of claim terms. For example, if the applicant described
his or her invention as including something or not including something then the claims are to be
interpreted accordingly. The basic principle, as I understand it, is that the claims are to be
interpreted consistently with applicant’s description of his or her invention during prosecution.
C.

Means-plus-function claiming.

64.

I understand that a patentee may express a claim limitation as a means for

performing a given function without specifying the structure that performs the function. These
kinds of terms are called “means-plus-function” terms, and are subject to special rules of claim
construction.
65.

Essentially, a structural limitation says what something is, while a functional

limitation says what it does. I understand that one determines whether a claim term is in meansplus-function form by considering whether the claims disclose a structure, or whether the
limitation speaks in purely functional terms. A term that uses words like “means for” is
presumed to be in means-plus-function form, while a term that does not use similar “means”
phrasing is presumed not to be in means-plus-function form. Those are only presumptions,
however, and the true test is whether the claim recites sufficient structure. I am advised that, as
- 27 -

A/74514379.11/3005005-0000352152

an aid in determining whether sufficient structure is recited in the claim, the Federal Circuit has
considered “whether the ‘term, as the name for a structure, has a reasonably well understood
meaning in the art.’”
66.

It is my understanding that means-plus-function limitations are construed to

“cover the corresponding structure, material, or acts described in the specification and
equivalents thereof.” 35 U.S.C. § 112, ¶ 6. Means-plus-function limitations are construed using
a two-step process. First, the court uses ordinary principles of claim construction to determine
the function explicitly set forth in the claims. Second, the court determines from the perspective
of a person of ordinary skill in the art, what structure, if any, disclosed in the specification
corresponds to the claimed function. Any such structure must be clearly associated with the
performance of the function.
D.

Indefiniteness

67.

I also have been informed on legal principles concerning patent validity, including

the definiteness requirement, and have used those principles in forming my opinions.
68.

I understand that in order for a patent to be valid, it must satisfy the statutory

definiteness requirement. The definiteness requirement provides that the patent specification
must conclude with one or more claims particularly pointing out and distinctly claiming the
subject matter that the applicant regards as the invention. In determining whether a patent claim
is indefinite, a court must consider whether one skilled in the art would understand the bounds of
the claim when read in light of the specification. Whether one skilled in the art would
understand the claim is dependent on whether the claim can be construed to the extent necessary
to resolve any dispute between the parties.
69.

In the case of means-plus-function limitations, it is my understanding that if a

corresponding structure for performing the function is not set out in the specification, then the
claim is indefinite.
- 28 -

A/74514379.11/3005005-0000352152

V.

SEVERAL TERMS IN THE PATENTS-IN-SUIT ARE INDEFINITE
A.

The “Reference Symmetry” Limitations are Indefinite.

70.

Claim 1 of the ‘988 patent requires: “recording multiple readings of the Wi-Fi

access point at different locations around the Wi-Fi access point so that the multiple readings
have reference symmetry relative to other Wi-Fi access points in the target area.”
71.

Claim 1 of the ‘694 patent requires that “the database records for substantially all

Wi-Fi access points in the target area provide reference symmetry within the target area.” It is
my opinion that both “reference symmetry” limitations are indefinite.
72.

“Reference symmetry” is not a term that has an established meaning in the art. I

have therefore looked to the specifications and prosecution histories for any information they
might have about what “reference symmetry” means.
73.

The patents describe “reference symmetry” in a context where there is significant

error in a location calculation caused by either too few reference points or by reference points
that lack balance or symmetry around the user. See Ex. C at 1:53-57, 2:53-57 (describing “[t]he
classic example” of a situation lacking reference symmetry as “along the shoreline” where there
are no access points in the water); Fig. 5 (entitled “Lack of Reference Symmetry” and showing
access points on only one side of a user); id. at 9:55-10:4 (describing Chinese Postman routing
resulting in a situation “in which there are numerous access point locations [602] on all sides of
the user [601] within the range [604] of the device’s 802.11 radio”); and Fig. 6 (showing access
points distributed around the user); id. at 9:49-10:4 (section entitled “Reference Symmetry”); Ex.
G at GSHFED188. The patent specification and figures thus seem to suggest that “reference
symmetry” has some relationship to a “balanced” or symmetrical distribution of numerous Wi-Fi
access points on all sides of the user device and within range of the user device’s Wi-Fi radio.
What that means in practical terms, however, is unclear.

- 29 -

A/74514379.11/3005005-0000352152

74.

Unless the target region is intentionally, and densely, seeded with access points,

there is simply no way to know whether reference symmetry has been provided in a given target
area. Symmetry of access points relative to a user is a function of the actual locations of the
access points relative to a user, and one-of-skill-in-the-art would have no way of knowing
whether a condition of “reference symmetry” does or does not exist. None of the asserted
patents contain any mention of intentional seeding. Access points may be distributed more or
less evenly, or in random and uneven ways. For example, a multi-tenant residential building
may have hundreds of access points on one block and there may be no other access points for
several blocks. Moreover, the patent would not even teach one of skill in the art when an
objective measure of the density or symmetry of access point deployment is such that “reference
symmetry” is achieved. Thus, there is a fundamental lack of any objective standard for
determining when the distribution of Wi-Fi access points might have reference symmetry with
regard to a user.
75.

Importantly, the claims do not require that Wi-Fi access points might have

reference symmetry with regard to a user—the only kind of reference symmetry discussed in the
specification. In the claims, there is no point of reference.
76.

In claim 1 of the ‘988 patent, it is the raw scanning data that must have “reference

symmetry” and the data must have it relative to other Wi-Fi access points in the target area. ‘988
patent, claim 1 (“calculated position information is obtained from recording multiple readings of
the Wi-Fi access point at different locations around the Wi-Fi access point so that the multiple
readings have reference symmetry relative to other Wi-Fi access points in the target area”)
(emphasis added). In claim 1 of the ‘694 patent, “the database records for substantially all Wi-Fi
access points in the target area provide reference symmetry within the target area.” The claim
language is vague and uncertain because there is no workable frame of reference for the
symmetry. In the ‘988 patent, the symmetry apparently involves measurements around each Wi- 30 -

A/74514379.11/3005005-0000352152

Fi access point relative to some other Wi-Fi access points. It is not clear what the relationship is
or what the symmetry looks like.
77.

In claim 1 of the ‘694 patent, “the database records for substantially all Wi-Fi

access points in the target area provide reference symmetry within the target area.” The claims
do not identify what within the target area is provided with “reference symmetry” by the
database records. Nor is it clear when the database provides such reference symmetry, how to
measure it, and how to distinguish it from prior art databases or allegedly infringing databases.
There is no objective measure for the analysis.
B.

“Arterial bias” and “avoid(s) arterial bias.”

78.

The inventors appear to believe that their proposed database creation technique of

deliberately taking measurements along every street will reduce a source of Wi-Fi device
location error that they refer to as “arterial bias.” I do not believe that this term has a recognized
technical meaning. Rather, I believe the meaning of “arterial bias” and “avoid arterial bias”/
“avoids arterial bias” found in the claims must be obtained from reading the patent and
prosecution history.
79.

Based on my review of the specification and prosecution history, it is my opinion

that the term “arterial bias” standing alone can be understood to mean “the deviation of the
calculated position information for a Wi-Fi access point toward heavily trafficked roads and
away from the actual geographic location of the access point due to the tendency of random
scanning to result in a greater number of scans from heavily trafficked roads.”
80.

I note that Skyhook’s own proposed construction is similar, although it does not

agree with the phrase “due to the tendency of random scanning to result in a greater number of
scans from heavily trafficked roads.” That is an important part of how the inventors described
“arterial bias.” Both the specification and relevant prosecution history emphasize the inventors’
belief arterial bias results from using the Random Model of data collection. See Ex. C at
- 31 -

A/74514379.11/3005005-0000352152

Abstract, 3:12-18, 4:4-9, 5:24-37, 7:55-63, 8:6-13, 8:24-27, 9:57-61, Figs. 3 & 5; Ex. G at
GSHFED00000187-88. The inventors’ solution to the problem of arterial bias is the method of
scanning to collect Wi-Fi access points by driving along a Chinese Postman route. This avoids
random data collection and unsystematic data collection that the Skyhook patents criticize.
81.

The claim language in claim 1 of the ‘988 and ‘694 patents “avoid[s] arterial bias”

does not, however, appear to me to be capable of any sort of objective determination. Avoidance
of arterial bias appears to be a relative measure; it is not clear what it is relative to. Based on the
claims and specification, one of skill-in-the-art would not know whether a particular database
creation technique, or a Wi-Fi device location calculation, will in fact have achieved such
“avoidance” within a given target area.
82.

There are several reasons for the difficulty in resolving this term and knowing

whether any particular database meets this term: (i) there is no criteria for measuring the degree
of bias toward streets in the target area; (ii) Chinese postman itself includes driving certain
streets more than once and therefore shows the very collection bias sought to be avoided (see
Figure 4 of the ‘988, ‘645 and ‘245 patents, showing driving every perimeter street twice); and
(iii) many access points in the target area may have a street on only one or two sides thus biasing
measurements to those streets. Accordingly, merely driving a measurement vehicle along every
street in a deliberate fashion does not necessarily “avoid arterial bias” and it is unclear whether
or to what extent this requirement is met or could be objectively measured in any given target
area. The determination might involve the efficiency of the route followed, the number of
measurements taken on each street segment, speed of the vehicle during scanning, the location of
each access point relative to streets, and the comprehensiveness of the scanning, to list some of
the factors to consider. Even with all of this information, it is unclear how one would use it to
determine whether or not there is an arterial bias. As such, it is my opinion that the claims in
which the phrase “avoid arterial bias” appears are indefinite; one of skill-in-the art simply would
- 32 -

A/74514379.11/3005005-0000352152

not know whether a particular system or method actually “avoids arterial bias” in a given target
area.
83.

If it means anything, “avoids arterial bias” means “eliminate arterial bias.” The

clearest ordinary meaning of “avoid” is “eliminate.” See, e.g., Exhibit P at GSHFED4446.
However, even this definition is problematic, for the reasons discussed above. Because each
recorded location of an access point always coincides with an actual GPS location of the
measuring vehicle, the measurements may be skewed toward the locations at which
measurements are actually made—toward the arteries and streets of the target area. Thus, if the
vehicle drives along only the streets within the service area, the estimated locations of the access
points contained in the database can be expected to be skewed toward those streets. Since these
serve as the basis for the computed location of a Wi-Fi device, these computations will also be
skewed.
C.

The “Logic” Terms.

84.

Each asserted claim of the ‘988 patent recites limitations that require “logic” for

performing a particular recited function. Specifically, the claims of the ‘988 patent require:
•

“logic to recalculate position information for Wi-Fi access points previously
stored in the database to utilize position information for the newly-discovered
readings of previously stored Wi-Fi access points” (‘988/1);

•

“computer-implemented logic to add records to the database for newly-discovered
Wi-Fi access points” (‘988/1);

•

“computer-implemented clustering logic to identify position information based on
error prone GPS information” (‘988/2);

•

“logic to determine a weighted centroid position for all position information
reported for an access point” (‘988/3);
- 33 -

A/74514379.11/3005005-0000352152

•

“logic to identify position information that exceeds a statistically-based deviation
threshold amount away from the centroid position” (‘988/3); and

•

“the clustering logic … excludes such deviating position information from the
database and from influencing the calculated positions of the Wi-Fi access points”
(‘988/3).

85.

I am informed that in a means-plus-function claim in which the relevant structure

is a computer or microprocessor programmed to carry out an algorithm, the disclosed structure
has to be more than a general purpose computer. Rather, I understand that the corresponding
structure is the computer or microprocessor as programmed to carry out an algorithm. If the
specification does not disclose with sufficient specificity how the computer is programmed to
carry out the particular function stated in the claim, it is my understanding that there is then no
corresponding structure disclosed and the limitation should be considered indefinite.
86.

As discussed below, I have considered the claims, and it is my opinion that

“logic” is not a structure, and that these terms are therefore means-plus-function terms. I have
reviewed the disclosure of the ‘988 patent, and for the reasons discussed below, it is my opinion
that it does not disclose corresponding structures capable of performing the functions stated in
the “logic” limitations.
1.
87.

The “logic” terms are functional.

“Logic” is not a structural term. It is not clear from the claims what the relevant

“logic” actually is, only what it does. A person of ordinary skill in the art would understand
“logic” to mean a series of defined steps for performing a function as opposed to a structure.
The steps of the function cannot be determined simply by stating a concept that may be logical –
the steps themselves must be provided in order for a structure to be disclosed.

- 34 -

A/74514379.11/3005005-0000352152

2.
88.

The claims and specification of the ‘988 patent do not disclose any
corresponding structures for the “logic” terms.

I have looked at the various logic elements found in the claims of the ‘988 patent,

and find that there is not any structure such as an exemplary algorithm that is disclosed (in either
the claims themselves or in the specification) that performs the function corresponding to the
means. It is therefore my opinion, that each of the “logic” terms in the ‘988 patent are indefinite.
a.

89.

“logic to recalculate position information for Wi-Fi access
points previously stored in the database to utilize position
information for the newly-discovered readings of previously
stored Wi-Fi access points” (claim 1)

For example, in the case of

logic to recalculate position information for Wi-Fi access points
previously stored in the database to utilize position information for
the newly-discovered readings of previously stored Wi-Fi access
points

the recalculating function associated with “logic” means is italicized. The specification, however,
fails to clearly disclose an algorithm that performs this function. In its proposed construction,
presented in the table below, Skyhook appears to suggest that the corresponding structure is an
algorithm that “would include a weighting value based on the age of the records.” However, the
specification nowhere provides an example of how this is implemented and in particular says
nothing about how the existing access points are repositioned or how the weighted average
model is built. My understanding, therefore, is that the structure has not been sufficiently
disclosed.
90.

It is my understanding that the parties have exchanged constructions and taken the

following positions with regard to this limitation:

- 35 -

A/74514379.11/3005005-0000352152

Google’s Position
This limitation is indefinite in violation of 35
U.S.C. § 112, ¶ 2 because the specification
does not disclose a structure corresponding to
the claimed “logic” capable of performing the
recited function of “recalculat[ing] position
information for Wi-Fi access points previously
stored in the database to utilize position
information for the newly-discovered readings
of previously stored Wi-Fi access points.”

Skyhook’s Position
Software and/or hardware to recalculate
position information for Wi-Fi access points
previously stored in the database. This
recalculation utilizes new position information
for such Wi-Fi access points calculated using
scans taken after the previously stored Wi-Fi
access points were stored.
See “Wi-Fi access points.”
This is not a means plus function claim
element. If the Court were to construe this
claim element as a means plus function claim
element, then Skyhook identifies the following
corresponding structure:
12:24–38
“An additional enhancement to the
algorithm would include a weighting value
based on the age of the records such that new
records represent a more significant indication
of the present location for a given access
point.
Once the parsing process has been
completed the central network system . . .
begins processing the new data. . . . 2) existing
access points are repositioned based on any
new data recorded by the scanners. The . . .
algorithm factors in the number of records and
their associated signal strengths to weight
stronger signal readings more than weaker
signals with a quasi weighted average model.”

91.

Skyhook’s proposed structure is not “logic” capable of recalculating the positions

of a previously stored Wi-Fi access point based on a new reading of those same Wi-Fi access
point. The description given is too general, and as noted below, contradicts other descriptions of
the process given in the specification. Ex. C at 12:31-38. In addition, although there is a set of
equations in columns 12 and 13 of the ‘988 patent that appear to describe calculations for
estimating positions for a number of readings, the specification does not provide an explanation
of how to combine these calculations with the requirements identified by Skyhook as the
required structure.

- 36 -

A/74514379.11/3005005-0000352152

92.

The specification presents, at best, contradictory information regarding what steps

would need to be carried out to perform such “logic.” The technique to avoid erroneous data in
determining the Wi-Fi positions and the technique to use newly-discovered position
information to improve the quality of previously gathered and determined position information
introduced in the ‘988 specification (Ex. C) at 5:37-41 are in contradiction to each other. With
regards to avoiding erroneous data, the specification explains that erroneous information is
“removed by the filter” “if 90% of the readings are within 200 meters of each other but the
remaining 10% of the readings are 5 kilometers” and “the centroid is recalculated with the
remaining location records,” Ex. C at 12:8-19. The specification attempts to reconcile the
addition of “newly-discovered position information” by explaining that “the error records may be
the result of an access point that has moved,” and confusingly explains that “[i]n this instance,
the centroid for the access points will quickly ‘snap’ to the new location based on the
preponderance of records,” Ex. C at 12:21-28.
93.

However, these statements run contrary to common sense. First, if records for

access points that are moved farther than a threshold distance are filtered out of a calculation, it
is unclear how the calculated centroid would move when the new recordings are not considered
in the calculation. In addition, one of skill in the art would assume that the “preponderance of
records” would favor an older location rather than a new location—presumably an access point
that has remained in one place for some time and then suddenly moved would have more records
in its original location than in its new location because the newer the location, the fewer the
readings associated with that location. This is acknowledged by the specification, which
identified problems in prior art databases where “the data across the databases are not
contemporaneous; some of the data is new while other portions are 3-4 years old. The age of the
access point location is important since over time access points can be moved or taken offline.”
Ex. C at 3:3-6.

- 37 -

A/74514379.11/3005005-0000352152

94.

The specification attempts to resolve this ambiguity by adding the descriptions

claimed by Skyhook to disclose the necessary structure. First, the specification states “[a]n
additional enhancement to the algorithm [that] would include a weighting value based on the age
of the records such that new records represent a more significant indication of the present
location for a given access point.” Ex. C at 12:21-28. Unfortunately, no specific value is given
for the appropriate weight that would be assigned to each record based on age, even if the
information was not filtered out of the data as described earlier in the specification. Next, the
specification goes on to describe an algorithm where “existing access points are repositioned
based on any new data recorded by the scanners” by “factor[ing] in the number of records and
their associated signal strengths to weight stronger signal readings more than weaker signals with
a quasi weighted average model,” Ex. C at 12:29-38. However, the specification does not
explain, for example, how to ensure that new readings with weak signal strengths for an access
point would interact with old readings with strong signal strength when determining position—
there is no “logic” described, merely disconnected factors. Nor does the specification explain the
logical differences between a quasi weighted average model and the criticized rudimentary
weighted average model.
95.

Finally, the specification discloses that the algorithm would include a “weighting

value” but does not disclose how the records would actually be recalculated. It says nothing
about how the existing access points are repositioned, or how the weighted average model is
built. Simply stating what an algorithm may include does not sufficiently disclose it in such a
manner that one skilled in the art would know and understand what structure corresponds to the
limitation.
96.

Furthermore, Skyhook asserts that the construction should be “[s]oftware and/or

hardware to recalculate position information for Wi-Fi access points previously stored in the
database.” However, as stated above, the disclosed structure must include the claimed algorithm.
- 38 -

A/74514379.11/3005005-0000352152

Skyhook has not identified in its proposed construction or in its cited corresponding structure the
algorithm for which the software/hardware is programmed.
b.
97.

“computer-implemented logic to add records to the database
for newly-discovered Wi-Fi access points” (claim 1)

In the case of
computer-implemented logic to add records to the database for
newly-discovered Wi-Fi access points

the recalculating function associated with “logic” means is italicized. The specification, however,
fails to clearly disclose an algorithm that performs this function. Skyhook appears to suggest
that the corresponding structure includes an algorithm that “factors in the number of records and
their associated signal strengths to weight stronger signal readings more than weaker signals with
a quasi weighted average model.” However, the specification nowhere provides an example of
how this is implemented, how the formula is adjusted for the number of records, or how the
weighted average model is built. My understanding, therefore, is that the structure has not been
sufficiently disclosed.
98.

It is my understanding that the parties have taken the following positions with

regard to this limitation:
Google’s Position
This limitation is indefinite in violation of 35
U.S.C. § 112, ¶ 2 because the specification
does not disclose a structure corresponding to
the claimed “computer-implemented logic”
capable of performing the recited function of
“add[ing] records to the database for newlydiscovered Wi-Fi access points.”

Skyhook’s Position
Computer-implemented software and/or
hardware to add data records to the database
for newly-discovered Wi-Fi access points.
See “Wi-Fi access points” and “database
records.”
This is not a means plus function claim
element. If the Court were to construe this
claim element as a means plus function claim
element, then Skyhook identifies the following
corresponding structure:
12:29–38
“Once the parsing process has been completed
- 39 -

A/74514379.11/3005005-0000352152

Google’s Position

99.

Skyhook’s Position
the central network system . . . begins
processing the new data. During this process
1) new access points are added to the database
and their physical location is calculated . . . .
The . . . algorithm factors in the number of
records and their associated signal strengths to
weight stronger signal readings more than
weaker signals with a quasi weighted average
model.”

The recited function is “add[ing] records to the database for newly-discovered

Wi-Fi access points.” Skyhook contends that the corresponding structure includes an algorithm
that “factors in the number of records and their associated signal strengths”; however, the
specification nowhere provides any further details on how the algorithm accounts for changes in
the numbers of records or signal strengths. The specification discloses that the algorithm would
include a “quasi weighted average model” (discussed below) but does not disclose how the
records would actually be added. It lacks information about how each new access point is added
or used in a weighted average.
100.

Furthermore, Skyhook asserts that the construction should be “[c]omputer-

implemented software and/or hardware to add data records to the database for newly-discovered
Wi-Fi access points.” However, the disclosed structure must include the claimed algorithm.
Skyhook has not, in its proposed construction or in its cited corresponding structure, identified a
definite algorithm for which the software/hardware is programmed.
101.

The specification criticizes prior art databases that “include the calculated position

of scanned access points rather than the raw scanning data obtained by the 802.11 hardware”
where each of the databases “calculates the access point location differently and each with a
rudimentary weighted average formula,” Ex. C at 3:18-22. However, the explanation that “new
access points are added to the database and their physical location is calculated,” and that the
- 40 -

A/74514379.11/3005005-0000352152

“algorithm factors in the number of records and their associated signal strengths to weight
stronger signal readings more than weaker signals with a quasi weighted average model” is not
specific enough to inform a person of ordinary skill in the art of the differences between the
formula criticized by the patentees and the proposed method. Ex. C at 12:32-38. The
specification does not explain why a “rudimentary weighted average formula is distinct from the
“quasi-weighted average model” disparaged a few paragraphs earlier and in particular does not
provide an explanation of the logical differences between them. The specification also describes
a calculated, final {Lati, Longi}, a “calculation” that “is then used as the final centroid value for
the location of that access point” and is “stored in the database,” but does not distinguish why
this storage of this calculation is an improvement over the calculations disparaged in the prior
art. Ex. C at 13:26-30, cf. 3:18-22. There is therefore nothing in the specification, nor in the art,
that explains in any definite way logic intended by the inventors for adding newly discovered
access points to the claimed database.
c.
102.

“computer-implemented clustering logic to identify position
information based on error prone GPS information” (claim 2)

In the case of
computer-implemented clustering logic to identify position
information based on error prone GPS information

the recalculating function associated with “logic” means is italicized. The specification, however,
fails to clearly disclose an algorithm that performs this function. Skyhook appears to suggest
that the corresponding structure includes an algorithm that “uses a definable threshold based on
the sigma of this distribution to filter out access points that are in error.” However, other than
one very specific example the specification nowhere provides an example of how this is
implemented, and in fact contains contradictory descriptions of how this process would be
carried out. My understanding, therefore, is that the structure has not been sufficiently disclosed.

- 41 -

A/74514379.11/3005005-0000352152

103.

It is my understanding that the parties have taken the following positions with

regard to this limitation:
Google’s Position
This limitation is indefinite in violation of 35
U.S.C. § 112, ¶ 2 because (1) it does not
apprise one skilled in the art of the bounds of
the claim, and in particular fails to apprise the
person of ordinary skill in the art what
constitutes “error prone GPS information”;
and (2) in violation of 35 U.S.C. § 112, ¶ 2
because the specification does not disclose a
structure corresponding to the claimed
“computer-implemented clustering logic”
capable of performing the recited function of
“identify[ing] position information based on
error prone GPS information.”

Skyhook’s Position
Computer-implemented software and/or
hardware to identify when position
information for a Wi-Fi access point based on
GPS readings is likely to be erroneous. The
software and/or hardware identifies position
information that is not located within a certain
threshold distance of other position
information for the Wi-Fi access point.
This is not a means plus function claim
element.
If the Court were to construe this claim
element as a means plus function claim
element, then Skyhook identifies the following
corresponding structure:
12:1-12:10
“In some cases the GPS receiver may record
erroneous or error records for some period of
time, which could negatively affect the final
access point location calculation. The parser
and filter process identifies these bad records
and either corrects them or removes them from
the system. The filtering process users
clustering techniques to weed out error prone
GPS readings. For example, if 90% of the
readings are within 200 meters of each other
but the remaining 10% of the readings are 5
kilometers away then those outliers are
removed by the filter . . . .”

104.

The recited function is “identify[ing] position information based on error prone

GPS information.” The specification fails to clearly link corresponding structure to the recited
recalculating position information function. Although Skyhook contends that the corresponding
structure is a filtering process that uses “clustering techniques to weed out error prone GPS
readings,” the specification does not provide further details anywhere. The specification

- 42 -

A/74514379.11/3005005-0000352152

discloses that the algorithm uses “clustering techniques,” but does not disclose what that
technique is. It only refers to an exemplary concept.
105.

Further, Skyhook asserts that the construction should be “[c]omputer-

implemented software and/or hardware to identify when position information for a Wi-Fi access
point based on GPS readings is likely to be erroneous.” As with its other constructions of this
type, Skyhook has not identified the specific algorithm in the specification for which the
software/hardware is programmed in its proposed construction and therefore has not identified
structure corresponding to the claimed “logic.”
106.

With regards to “computer-implemented clustering logic to identify position

information based on error prone GPS information” limitation (found in claim 2 of the ‘988
patent) and “the clustering logic … excludes such deviating position information from the
database and from influencing the calculated positions of the Wi-Fi access points” limitation
discussed below (found in claim 3 of the ‘988 patent) among other ambiguities, in the only place
outside of the claims or the summary where the specification discusses “clustering,” the
specification describes “calculat[ing] the weighted centroid for the access point using all reported
data . . . then determin[ing] the standard deviation based on the distribution of the reported
locations” and “us[ing] a definable threshold based on the sigma of this distribution to filter out
access points that are in error.” Ex. C at 12:14-20. However, only one example of a “threshold”
is given “if 90% of the readings are within 200 meters of each other but the remaining 10% of
the readings are 5 kilometers away then those outliers are removed by the filter.” Ex. C at 12:611. The specification fails to define what threshold, other than the specific example given, would
indicate access points that are in error.
107.

In addition, as discussed above with respect to the “logic to recalculate position

information for Wi-Fi access points previously stored in the database to utilize position
information for the newly-discovered readings of previously stored Wi-Fi access points”
- 43 -

A/74514379.11/3005005-0000352152

limitation found in claim 1 of the ‘988 patent, the specification fails to describe any logic for
how new data on “moved” access points would be accounted for in instances where an access
point moves outside a threshold distance. Furthermore, the patent fails to identify what errors
are caused by GPS errors as opposed to other types of errors. See generally Ex. C at 12:1-20, cf.
12:21-28 (“[n]ote that the error records may be the result of an access point that has moved”).
d.

108.

“the clustering logic … excludes such deviating position
information from the database and from influencing the
calculated positions of the Wi-Fi access points” (claim 3)
Skyhook does not propose a construction for this term. The specification

generalizes the concept in a summary. Ex. C at 12:16-17 (“For example, if 90% of the readings
are within 200 meters of each other but the remaining 10% of the readings are 5 kilometers away
then those outliers are removed by the filter . . . The system uses a definable threshold based on
the sigma of this distribution to filter out access points that are in error. Once these error records
are marked, the centroid is recalculated with the remaining location records to determine the
final centroid.”). The key points of the concept are not disclosed, such as how the threshold,
sigma, and centroid are calculated.
109.

Therefore, because the means-plus-function “logic” claims of the ‘988 patent lack

corresponding structures, they are indefinite.
e.
110.

“logic to determine a weighted centroid position for all position
information reported for an access point” (claim 3)

In the case of
logic to determine a weighted centroid position for all position
information reported for an access point

the recalculating function associated with “logic” means is italicized. The specification, however,
fails to clearly disclose an algorithm that performs this function. Skyhook appears to suggest
that the corresponding structure includes an algorithm that “calculates the weighted centroid for
- 44 -

A/74514379.11/3005005-0000352152

the access point using all reported data.” However, the specification nowhere provides an
example of how this is implemented. My understanding, therefore, is that the structure has not
been sufficiently disclosed.
111.

It is my understanding that the parties have taken the following positions with

regard to this limitation:
Google’s Position
This limitation is indefinite in violation of 35
U.S.C. § 112, ¶ 2 because the specification
does not disclose a structure corresponding to
the claimed “logic” capable of performing the
recited function of “determine[ing] a weighted
centroid position for all position information
reported for an access point.”

Skyhook’s Position
Software and/or hardware to determine a
weighted centroid position for a Wi-Fi access
point. The weighted centroid position is
determined using all position information
reported for that Wi-Fi access point.
See “weighted centroid position” and “Wi-Fi
access points.”
This is not a means plus function claim
element.
If the Court were to construe this claim
element as a means plus function claim
element, then Skyhook identifies the following
corresponding structure:
12:11–13
“In particular, the system first calculates the
weighted centroid for the access point using
all reported data.”
12:34–44
“The . . . algorithm factors in the number of
records and their associated signal strengths to
weight stronger signal readings more than
weaker signals with a quasi weighted average
model.
During data gathering, a WPS user is
equipped with a Wi-Fi receiver device which
measures Received Signal Strength (RSS)
from all the available Wi-Fi access points, and
then extracts location information of
corresponding access points.”

- 45 -

A/74514379.11/3005005-0000352152

112.

The recited function is “determine a weighted centroid position for all position

information reported for an access point.” The specification fails once again to clearly link
corresponding structure to the claimed recalculating position information function. Skyhook
contends that the corresponding structure is an algorithm that “factors in the number of records
and their associated signal strengths to weight stronger signal readings more than weaker signals
with a quasi weighted average model.” But while it is true that the specification discloses that
the “system first calculates the weighted centroid for the access point using all reported data,”
critically, it does not disclose what that calculation is. As discussed above at paragraph 87,
simply naming a conceptual algorithm does not sufficiently disclose the steps of the algorithm in
such a manner so as to disclose a structure.
113.

In particular, despite the applicants’ initial criticism that “each of the databases

calculates the access point location “with a rudimentary weighted average formula,” Ex. C at
3:18-22, they do not offer any alternative. The specification of the ‘988 patent states that “the
system first calculates the weighted centroid for the access point using all reported data” without
any other detail about how the weighted centroid position is calculated, Ex. C at 12:11-13, failing
to distinguish it from a “rudimentary” weighted average formula. The specification then states
that “the centroid is recalculated with the remaining location records to determine the final
centroid,” with no further explanation of the initial weighted centroid.. As with the “computerimplemented logic to add records to the database for newly-discovered Wi-Fi access points”
limitation found in claim 1 of the ‘988 patent, the inventors fail to explain why a “rudimentary
weighted average formula” is distinct from a “quasi-weighted average model,” and in particular
the specification provides no explanation of the logical differences between them. There is

- 46 -

A/74514379.11/3005005-0000352152

therefore nothing in the specification, nor in the art, that explains what is meant by “logic to
determine a weighted centroid position for all position information reported for an access point.”
114.

Further, Skyhook asserts that the construction should be “[s]oftware and/or

hardware to determine a weighted centroid position for a Wi-Fi access point.” Skyhook has not
identified the specific algorithm for which the software/hardware is programmed in its proposed
construction.
f.

115.

“logic to identify position information that exceeds a
statistically-based deviation threshold amount away from the
centroid position” (claim 3)
Finally, in the case of
logic to identify position information that exceeds a statisticallybased deviation threshold amount away from the centroid position

the recalculating function associated with “logic” means is italicized. The specification, however,
fails to clearly disclose an algorithm that performs this function. Skyhook appears to suggest
that that the corresponding structure includes the same algorithm that it earlier describes for
“identify position information based on error prone GPS information,” namely “uses a definable
threshold based on the sigma of this distribution to filter out access points that are in error.”
However, other than that one very specific example referenced above at paragraph 106, the
specification nowhere provides an example of how this is implemented, and in fact contains
contradictory descriptions of how this process would be carried out. My understanding,
therefore, is that the structure has not been sufficiently disclosed.
116.

It is my understanding that the parties have taken the following positions with

regard to this limitation:

- 47 -

A/74514379.11/3005005-0000352152

Google’s Position
This limitation is indefinite in violation of 35
U.S.C. § 112, ¶ 2 because the specification
does not disclose a structure corresponding to
the claimed “logic” capable of performing the
recited function of “identify[ing] position
information that exceeds a statistically-based
deviation threshold amount away from the
centroid position.”

Skyhook’s Position
The software and/or hardware excludes such
deviating position information from being
stored in the database of WiFi access points.
Such deviating position information is not
used to determine the calculated positions of
the Wi-Fi access points.
See “computer-implemented clustering logic .
. . “and “Wi-Fi access points.”
This is not a means plus function claim
element.
If the Court were to construe this claim
element as a means plus function claim
element, then Skyhook identifies the following
corresponding structure:
12:1–12:19
“In some cases the GPS receiver may record
erroneous or error records for some period of
time, which could negatively affect the final
access point location calculation. The parser
and filter process identifies these bad records
and either corrects them or removes them from
the system. The filtering process users
clustering techniques to weed out error prone
GPS readings. For example, if 90% of the
readings are within 200 meters of each other
but the remaining 10% of the readings are 5
kilometers away then those outliers are
removed by the filter . . . . In particular, the
system first calculates the weighted centroid
for the access point using all reported data. It
then determines the standard deviation based
on the distribution of the reported locations.
The system uses a definable threshold based
on the sigma of this distribution to filter out
access points that are in error. Once these error
records are marked, the centroid is
recalculated with the remaining location
records to determine the final centroid . . . .”

117.

The recited function is “identify[ing] position information that exceeds a

statistically-based deviation threshold amount away from the centroid position.” The
specification again fails to clearly link corresponding structure to the recited recalculating
- 48 -

A/74514379.11/3005005-0000352152

position information function. Skyhook contends that the corresponding structure is a filtering
process that employs “clustering techniques to weed out error prone GPS readings” by
calculating a weighted centroid, determining the standard deviation, and using a definable
threshold based on the sigma of the distribution of the reported locations. As explained above at
108, the specification again does not provide enough detail into calculating the weighted
centroid, the standard deviation, or the definable threshold to allow one skilled in the art to know
and understand what structure corresponds to the limitation.
118.

Skyhook also asserts here that the construction should be “software and/or

hardware [that] excludes such deviating position information from being stored in the database
of WiFi access points.” For the same reasons as for Skyhook’s similar constructions, explained
at 96, 100, 105, 114, Skyhook has failed to identify the specific algorithm for which the
software/hardware is programmed in its proposed construction.
D.

“Predefined Rules” and “Rules” (‘897 patent, claims 1 and 3)

119.

In claim 1 of the ‘897 patent, the inventors claim the step of “using the recorded

location information for each of the observed WiFi access points in conjunction with predefined
rules to determine whether an observed WiFi access point should be included or excluded from a
set of WiFi access points” (emphasis added).
120.

In dependent claim 3 of the ‘897 patent, the “predefined rules” are “rules to

determine a reference point and to compare the recorded location information for each of the
observed WiFi access points to the reference point.”
121.

In my opinion, these terms provide no real guidance or description to a person of

ordinary skill in the art, and a person with ordinary skill in the art cannot use these terms to
determine what the claims do or do not cover. In the context of a computer database, almost
- 49 -

A/74514379.11/3005005-0000352152

anything could be a “rule,” and it would necessarily be “predefined” prior to the relevant
operation being executed.
122.

Although the ‘897 patent refers to clustering and filtering algorithms, see, e.g.,

Ex. F at 9:11-15, I understand that Skyhook contends these limitations should be construed more
broadly than the disclosed algorithms. In my opinion, these limitations do not clearly define
what is or is not within the scope of the claims.
E.

“Being suited” (‘245 patent, claim 1)

Claim 1 of the ‘245 patent requires “based on the number of Wi-Fi access points
identified via received messages, choosing a corresponding location-determination algorithm
from a plurality of location-determination algorithms, said chosen algorithm being suited for the
number of identified Wi-Fi access points.”
123.

The term “being suited” as used in the ‘245 patent I find to be ambiguous and

subjective. It is applied apparently to different types of equations or algorithms that can be used
for different numbers of access points. But the patent does not teach any objective way to
determine which equations are suitable and which are not, or even give particular criteria by
which suitability might be judged. And, given that the patent is supposed to be teaching one of
skill in the art how to determine location, the ambiguity here goes to the thrust of what should
have been provided in the patent. Accordingly, I find this term subjective and ambiguous. In my
opinion, it is indefinite.
VI.

THE PROPER CONSTRUCTION OF CERTAIN DISPUTED CLAIM TERMS
A.

“Target area”

124.

Claim 1 in each of the ‘988 and ‘694 patents requires “a database of Wi-Fi access

points for at least one target area having a radius on the order of tens of miles” and the database
must “includ[e] database records for substantially all Wi-Fi access points in the target area.”

- 50 -

A/74514379.11/3005005-0000352152

Claim 1 of the ‘245 patent similarly requires “a reference database of calculated locations of WiFi access points in a target area.”
125.

It is unsurprising that the majority of the claims require a “target area.” As

discussed above, the common specification is dominated by a single method for creating the
database, which involves driving a vehicle in a systematic manner along every street. The ‘988
patent explains that following the most efficient planned route through the entire target area is
important not only to acquire as many access points as possible, but also to achieve the
inventors’ key goals of avoiding arterial bias and achieving reference symmetry.
126.

The necessary first step in programmatically gathering information about Wi-Fi

access points is to identify the “target area” in which to plan a scanning route. Ex. C at 8:41-44
(“Preferred embodiments of the invention include a methodology for identifying a target region
for coverage and then using the Chinese Postman routing algorithm for planning the vehicle
route.”) & id. at 8:28-59; Ex. G at GSHFED187-189; Ex. H at GSHFED297-298. Without a preidentified “target area,” it would be impossible to “plan[] the vehicle route,” Ex. C at 8:43-44, or
“perform[] a planned audit … following designated scanning routes,” Ex. G at GSHFED189
(also discussing a “systematic” traverse of the target area). In simple terms, you cannot plan a
route if you do not know where you are going.
127.

The patent and prosecution history also confirm that once the relevant geographic

region is identified, the shortest possible route must be pre-planned for traversing all drivable
roads. See, e.g., Ex. C at 7:37-44 (“These vehicles 201 follow a programmatic route through
target scan areas to gather data in the most optimal fashion producing the highest quality data.
The target scan areas typically represent a large metropolitan area including every single drivable
street in 15-20 mile radius”) (emphasis added); Ex. G at GSHFED188 (criticizing failure in the
prior art to “follow[] designated scanning routes”); id. (“Collecting multiple readings of Wi-Fi

- 51 -

A/74514379.11/3005005-0000352152

access points in a systematic fashion, as described in the application, provides reference
symmetry within the target area.”); Ex. G at GSHFED187-89.
128.

In light of the specification and the prosecution history, it is my opinion that

“target area” should be defined as “a pre-identified geographic region throughout which a
shortest route is planned along all drivable roads.”
129.

I have considered Skyhook’s proposal that a “target area” should be defined as a

“targeted geographic area.” It is my opinion that Skyhook’s proposed definition is incorrect.
Skyhook’s definition provides no clarity whatsoever. Although a “target area” is clearly a
“geographic area,” that is not all that it is. As explained above, to receive any benefit from the
invention as described in the patent, the “target area” must be pro-actively identified in advance
of data collection so that the route can be planned. You cannot decide what the “target area” is
after scanning is complete. Skyhook’s construction also is in error because it disregards the
teaching of the patents and prosecution history that the claimed “target area” is an area that is
traversed efficiently, for example, by using the Chinese Postman method.
B.

The Location Terms: “calculated position information” (‘988/1 & ‘694/1);
“calculated positions of the Wi-Fi access points” (‘988/3); “calculated
locations” (‘245/1 & 2); “recorded location information” (‘897/1 & 3)

130.

In every patent, each of the asserted claims requires a database of calculated Wi-

Fi access point locations. The patents-in-suit use several phrases for the same concept:
“calculated position information” for the Wi-Fi access point (‘988/1 and ‘694/1), “calculated
positions of the Wi-Fi access points” (‘988/3), “calculated locations” (‘245/1 and 2), and
“recorded location information” (‘897/1 & 3) are all synonymous. It is my opinion that each of
these terms should be defined as:
The physical location (i.e., latitude and longitude) attributed to
each Wi-Fi access point determined mathematically from readings
recorded along a shortest planned route throughout all drivable
roads in the target area (i.e., by following the Chinese Postman
routing algorithm). The “calculated position information” cannot
- 52 -

A/74514379.11/3005005-0000352152

be based on randomly, or non-systematically, collected readings of
Wi-Fi access points.
In reaching this opinion, I have relied on the language of the claims themselves, the
specification, and the applicants’ own characterization of their invention in their correspondence
with the PTO during prosecution.
131.

These sources indicate that the calculated position of a Wi-Fi access point is its

estimated physical location as determined based on scanning data that was recorded. See Ex. C at
4:28-40. But they also indicate to a person of ordinary skill that the claimed “calculated position
information” cannot be based on just any scanning data.
132.

Instead, a person of ordinary skill would understand that the claimed calculated

position information must be based on data gathered while systematically and efficiently
traversing all drivable roads in the target area using a Chinese Postman routing algorithm in
order to avoid arterial bias, and cannot be based on Wi-Fi access point readings collected
randomly, or non-systematically. The inventors criticized the Random Model, and the problem
of arterial bias, that they believe results from it, extensively in their patent. Ex. C at 3:12-18,
7:52-8:27, 9:57-61; Figs. 3, 5. They also directly contrasted the Random Model with their own
solution:
Another approach is develop routing algorithms that include every
single street in the target area so as to avoid arterial bias in the
resulting collection of data thus producing a more reliable
positioning system for the end users. FIG. 4 describes an optimized
routing algorithm known as the Chinese Postman to calculate the
most efficient driving route for covering every single street in a
target area.
Ex. C at 8:28:34; see also id. at 7:37-44 (“These vehicles 201 follow a programmatic route
through target scan areas to gather data in the most optimal fashion producing the highest quality
data. The target scan areas typically represent a large metropolitan area including every single

- 53 -

A/74514379.11/3005005-0000352152

drivable street in 15-20 mile radius”), Ex. C at 9:12-19 (discussing advantages of Chinese
Postman), Ex. C at 9:64-10:4 (same).
133.

In forming my opinion regarding the proper construction of the “location” terms, I

have also considered the inventors’ statements to the PTO during prosecution of the ‘988 and
‘694 patents.3 I note that Skyhook emphasized to the Examiner that the prior art of record was
“completely silent regarding any method of determining the location of Wi-Fi access points” and
“silent regarding any particular route or scheme taken or used by the mobile station to gather
location information about wireless access points.” Ex. G at GSHFED185 & GSHFED186. As
they did in the specification, the applicants contrasted random, or, non-systematic, data
collection with their invention:
[N]one of the cited reference teach or suggest conducting an audit
of an area to build a reference database of the locations of Wi-Fi
access points in a target area so as to provide reference symmetry
and avoid arterial bias. … Namely, as described in the application,
the location data collected by the wardrivers is often inaccurate,
incomplete, and grows organically rather than being collected in a
systematic fashion to purposefully avoid arterial bias. See
Application at ¶¶ 15-17.
Ex. G at GSHFED187-88. The cited paragraphs 15-17 of the Application correspond to
unamended claims 1-3 as recited in the Summary of the ‘988 patent at column 4, lines 28-57.
See Ex. G at GSHFED234.
134.

I also think it quite relevant that the applicants emphasized their method of

“[c]ollecting multiple readings of Wi-Fi access points in a systematic fashion, as described in the

3

Although I have also reviewed the prosecution histories of the ‘245 and ‘897 patents, I did not identify
any discussion in those documents between the applicants and the PTO that is relevant to the meaning of
the “location” terms.

- 54 -

A/74514379.11/3005005-0000352152

application,” and the resulting improvement in the calculated location information. Ex. G at
GSHFED188 (emphasis added). The applicants described this as their “solution[]” to “the
arterial bias problem”: “by performing a planned audit, and avoiding arterial bias, applicants at
least achieve more complete information about access points in the target area, higher quality
estimates of access point locations, and reference symmetry.” Ex. H at GSHFED298 (emphasis
added).
135.

I also note that during prosecution of the ‘694 patent, the applicants stated they

were “amending claim 1 to more particularly recite characteristics of the calculated position
information.” Ex. H at GSHFED298. This would indicate to a person of ordinary skill in the art
that the nature of the claimed “calculated position information” depends on the inventor’s
comprehensive and systematic method of collecting the underlying data.
C.

“substantially all Wi-Fi access points in the target area”

136.

Both of the independent claims of the ‘988 and ‘694 patents require that the

database “includ[e] database records for substantially all Wi-Fi access points in the target area.”
See ‘988 patent, claim 1; ‘694 patent, claim 1. The applicants stressed that the point of their
purported invention was to capture as many access points as possible, and preferably all of them.
See, e.g., Ex. C at 8:44-47 (“The scanning vehicle [401] follows the optimal route according to
the algorithm … ensuring that all observable access points are detected and mapped by the
system.”); Ex. G at GSHFED188 (“[B]y performing a planned audit … applicants at least
achieve more complete information about access points in the target area[.]”). Based on the
specification, a person of skill in the art would understand that “substantially all” access points
are required to serve the purpose of “achiev[ing] more complete information about access points
in the target area.” While not a concrete numerical standard, this in essence means “all but an
insignificant number of Wi-Fi access points in the target area.”

- 55 -

A/74514379.11/3005005-0000352152

137.

Skyhook’s proposal does not conform to the applicants’ stated goal of collecting

“more complete information.” By rewriting the claim to insert “observed” into “substantially all
Wi-Fi access points so that it becomes “substantially all observed Wi-Fi access points,” Skyhook
is divorcing the methodology disclosed in the specification—collecting access points using a
planned route—from the intended result of the method—”to fully explore and cover the streets of
a target region” in order to gather more access points than would be achieved through random
collection. See, e.g., Ex. C at 5:24-37. Skyhook’s rewriting would exclude the purpose of the
disclosed collection method from the boundaries of the claim. For example, a “target area”
might be scanned and 500 Wi-Fi access points included in the database of the claims. Six
months later, there might now be 1,000 WiFi access points in that same target area. If the target
area had not been rescanned during that time, the database would still have 500 access points,
and would still have “substantially all observed Wi-Fi access points” simply because no
observations had been made during a time in which the actual number of Wi-Fi access points
present changed dramatically. Skyhook’s proposed construction would destroy the usefulness of
the allegedly improved location system. Ex. C at 5:24-27.
138.

Skyhook’s proposed construction differs from the specification in another

important way. As noted above, the specification discuses how to “ensur[e] that all observable
access points are detected and mapped by the system.” Ex. C at 8:44-47 (emphasis added).
Observable access points are different from observed access points. The inventor’s goal was to
ensure that as many access points as possible were included in the database because the more
access points it has, the better the system is likely to be at determining the location of mobile
devices. But the claim does not require that each and every access point in the target area be
included in the database. The claim phrase “substantially all” accounts for the possibility that
some de minimus number of WiFi access points could be excluded from the database, whether
because they are unobservable of for some other reason. (The number of unobservable access
points is likely to be quite small relative to the total number of access points in the target area. A
- 56 -

A/74514379.11/3005005-0000352152

Wi-Fi access point that cannot be seen by devices with Wi-Fi radios cannot be used by them, and
the purpose of an access point is to provide network connectivity.)
D.

“providing a reference database of calculated locations of Wi-Fi access points
in a target area”

139.

Claim 1 of the ‘245 patent requires “providing a reference database of calculated

locations of Wi-Fi access points in a target area.” Consistent with my opinion regarding the
proper construction of “calculated locations,” the database of the ‘245 patent will contain
information about the physical location (i.e., latitude and longitude) attributed to each Wi-Fi
access point determined mathematically from readings recorded along a shortest planned route
throughout all drivable roads in the target area (i.e., by following the Chinese Postman routing
algorithm), and cannot be based on randomly, or non-systematically, collected readings of Wi-Fi
access points. See above at ¶¶ 130 - 135. For the same reasons the calculation of the WiFi
access point location excludes randomly, or non-systematically, collected readings of Wi-Fi
access points, so too must “a reference database of calculated locations” exclude such readings.
140.

In my opinion, the proper construction of “providing a reference database of

calculated locations of Wi-Fi access points in a target area” is therefore: “A database that
contains calculated locations for all the Wi-Fi access points collected in the pre-identified target
area by scanning a shortest planned route along all drivable roads. The database does not include
information about Wi-Fi access points gathered using random or end-user based collection
methods.”
E.

“a WiFi-enabled device communicating with WiFi access points within range
of the WiFi-enabled device so that observed WiFi access points identify
themselves”

141.

A device may use either active or passive scanning to discover the existence of

nearby access points. An “active” scan is done by broadcasting multiple requests to any wireless
access points in range and recording the responses, which are typically a unique identifier such
as a MAC address. See, e.g., Ex. F at Fig. 2, Fig. 5, 6:53-59, 7:12-26. In contrast, a “passive”
- 57 -

A/74514379.11/3005005-0000352152

scan does not broadcast a request, but merely records data sent out by any nearby wireless access
points that are both in range and broadcasting. Although both forms of scanning take a fraction
of a second to complete, passive scanning takes longer because the scanning device must wait for
signals sent out by APs at a set rate, rather than prompting a beacon to send a reply. Therefore,
active scanning is about ten times as fast as passive scanning but uses more overhead (e.g.,
battery power, bandwidth); the response time is generally 10 ms for active versus 100 ms for
passive scanning. See, e.g., Ex. BB, Leary et al., “Wireless LAN Fundamentals: Mobility,” Jan.
9, 2004, available at http://www.ciscopress.com/articles/article.asp?p=102282&seqNum=2
142.

One of ordinary skill in the art would understand the term “a WiFi-enabled device

communicating with WiFi access points within range of the WiFi-enabled device so that
observed WiFi access points identify themselves” in claim 1 of the ‘897 patent to mean “a user
device having a Wi-Fi radio actively searching for Wi-Fi access points by transmitting a signal to
all Wi-Fi access points within range and receiving a response that includes a unique identifier
(e.g., a MAC address) from each such Wi-Fi access point.”
143.

Skyhook asserts that the claim term does not require active searching for access

points, but rather encompasses passive scanning as well. I disagree. The language of the claim
itself specifically says the Wi-Fi access points identify themselves in response to a
communication from a Wi-Fi enabled device: “a WiFi-enabled device communicating with WiFi
access points within range of the WiFi-enabled device so that observed WiFi access points
identify themselves.” ‘897, claim 1 (emphasis added). This is the explicit definition of active
scanning under the IEEE 802.11 standard (a standard that is expressly cited in the patent itself);
see also id. at 7:15-19 (“The 802.11 radio sends out a probe request to all 802.11 access points
[204] within range. According to the 802.11 protocol, those access points in receipt of a probe
request will transmit a broadcast beacon containing information about the access point.”); Fig. 2.
A person of ordinary skill in the art would understand the act of communicating with an access
- 58 -

A/74514379.11/3005005-0000352152

point so that it responds to be the sending a probe request and receiving a response as disclosed
in the patent, and discussed in the 802.11 standard.
F.
144.

“in response to a user application request to determine a location of a userdevice having a Wi-Fi radio”
The second step of clam in the ‘245 patent is triggering a request to all Wi-Fi

access points within range “in response to a user application request to determine a location of a
user-device having a Wi-Fi radio.” I understand that Skyhook contends an operating system is a
“user application” within the meaning of the ‘245 patent. I do not agree.
145.

An operating system is not a user application—user applications run on an

operating system. See, e.g., the definitions of “application,” “utility”/”utility program,” and
“operating system” in the Microsoft Computer Dictionary (Ex. R at 31, 378, 544) and Webster’s
New World Computer Dictionary (Ex. S at 23, 264-265, 387). The ‘245 patent itself does not list
operating systems among applications. Consistent with this well-understood meaning in the art,
the exemplary applications useful on mobile devices listed in the ‘245 patent are all applications
that run on an operating system: “[m]obile email, walkie-talkie services, multi-player gaming
and call following,” keeping track of individuals or vehicles, and finding specific stores or
services near to the end user. See Ex. E at 1:31-42.
146.

The ‘245 patent also distinguishes between user applications, an operating system,

hardware, and the interfaces that support these. For example, Fig. 9, and the portion of the
specification that describes Fig. 9, describes separately user applications such as “an application
or service [901] that utilizes location readings,” an application programming interface (Fig. 9,
element 910) (commonly called an “API”), and other software including the “positioning
software 103” that interacts with the device hardware and other software to actually determine
the location as requested by the end-user facing application. Ex. E at 6:12-19; Fig. 1 (esp.
elements 101, 103); Fig. 9 (esp. elements 901, cf. 910, box 103).
147.

An API is an interface between different software programs, and is typically used

by an application program to communicate with the OS or some other program to direct the
- 59 -

A/74514379.11/3005005-0000352152

performance of another program. See, e.g., Ex. AA PC Magazine Encyclopedia (definition at
“API”) available at http://www.pcmag.com/encyclopedia_term/0,2542,t=API&i=37856,00.asp.
As this definition suggests, an operating system or program does not to need an API to mediate
interactions with itself. In the ‘245 patent, the API delivers location information to the “location
application.” Ex. E at 6:52-56; Fig. 9.
148.

A “user” or “end user” is commonly understood to be the consumer who is using

a device to perform tasks. See, e.g., the definitions of “end user” and “user” in the Microsoft
Computer Dictionary (Ex. R at 193) and Webster’s New World Computer Dictionary (Ex. S at
129 and 387). Furthermore, within the ‘245 patent, “users” are referred to as consumers or endusers of mobile devices. See Ex. E at 1:36-40 (parents, supervisors, and business travelers are
examples of “users” who “demand/seek applications” that are location aware), Ex. E at 8:27,
9:10, 9:54-55.
149.

In my opinion, it is therefore appropriate to construe the term “in response to a

user application request to determine a location of a user-device having a Wi-Fi radio” in claim 1
as “in response to a request made by an end-user facing application, i.e., not by the operating
system, to determine the location of an end user-device using a Wi-Fi radio.”
G.

“a user-device having a Wi-Fi radio”

150.

As described immediately above, the “users” referred to in the ‘245 patent are

consumers or end-users of mobile devices. See also, Ex. E at 1:36-40; 8:27; 9:10; 9:54-55.
Accordingly, the term “a user-device having a Wi-Fi radio” in claim 1 should be construed as
“an end-user or consumer device having a Wi-Fi radio.”
H.

“simple signal strength weighted average model”

151.

Claim 6 of the ‘245 requires that one of “the plurality of location determination

algorithms includes a simple signal strength weighted average model.” The specification
discloses determining a device’s location by “compar[ing] the list of observed access points
along with their calculated signal strengths to weight the location” of the device user (‘245
- 60 -

A/74514379.11/3005005-0000352152

patent, 7:1-3) and suggests several weighting methods including “triangulation techniques” and
“simple signal strength weighted average models” (see ‘245 patent 7:3-12). Weighting methods
based on signal strengths are described only once in the ‘245 specification, defined as taking the
average location of access point readings and using their associated signal strengths “to weight
stronger signal readings more than weaker signals.” See Ex. E at 12:29-32. See also Ex. T at
GSHFED_0004595 (“signal strength”), Ex. P at GSHFED_0004470 (“weighted average”), Ex.
U at GSHFED_0004431 (“weighted”). Specific mathematical calculations that would average
positions weighted by their associated signal strengths are shown in the ‘245 patent at 12:45-60
and 13:1-17; these are the only discussions of weighted average of signal strengths in the
specification.
152.

Although there is one mathematical formula in the specification for taking the

average of a location weighted by signal strength, it should be noted that this description in the
specification is made in the context of determining the position of a Wi-Fi access point—not in
the context of determining the location of a user device as in claim 6 of the ‘245 patent.
153.

Nevertheless, a person of ordinary skill in the art would understand a model to be

a mathematical formula, and a weighted average to be an average in which each component is
multiplied by a factor to give it its proper importance, where the sum of the factors equals one.
See Ex. T at GSHFED_0004594 (“model”), Ex. T at GSHFED_0004597 (“weighted average”),
Ex. P at GSHFED_0004448 (“to calculate” is “to . . .determine by arithmetical or mathematical
reckoning.”), Ex. P at GSHFED_0004470 (“weighted average”), Ex. U at GSHFED_0004431
(“weighted”), Ex. O at GSHFED_0004434 (“algebra”), Ex. V at GSHFED_0004442
(“calculate”), and Ex. W at GSHFED_0004479 (“algorithm”). Therefore a person of ordinary
skill in the art would understand the correct construction of “a simple signal strength weighted
average model” to be “a calculated average of signal strength measurements resulting from the
multiplication of each measurement by a numeric or algebraic factor to weight stronger signal
readings more than weaker signals, used to determine the location of a user device.”
- 61 -

A/74514379.11/3005005-0000352152

I.

“triangulation technique”

154.

Dependent Claim 8 of the ‘245 states: “The method of claim 1 wherein the

plurality of location-determination algorithms includes a triangulation technique.” The
specification describes a triangulation technique as one possible method “to weight the location
of” the device user. See Ex. E 6:67-7:12. The only “triangulation” described in the specification
are the techniques used to determine location would have been known to a person of ordinary
skill in the art at the time the application that became the ‘245 patent was filed. See generally
‘245 patent at 1:61–63 and 2:7–9, GSHFED_0004592 (“algorithm”), GSHFED_0004596
(“triangulation”), GSHFED_0004466-67 (“triangulate”), GSHFED_0004430 (“triangulate” and
“triangulation”).
155.

It is my opinion that a person of ordinary skill in the art would therefore

understand the construction of “triangulation technique” to be “calculating the physical location
of a user device by using the strength of signals received from two or more Wi-Fi access points
whose locations have been calculated, by the formation of triangles having the user device and
each such Wi-Fi access point as the vertices.”

- 62 -

A/74514379.11/3005005-0000352152