UNITED STATES PATENT AND TRADEMARK OFFICE
`
`____________
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`____________
`
`
`AMERICAN NATIONAL MANUFACTURING INC.,
`Petitioner,
`
`v.
`
`SLEEP NUMBER CORPORATION
`f/k/a SELECT COMFORT CORPORATION,
`Patent Owner.
`____________
`
`Case No. IPR2019-00500
`Patent No. 9,737,154 B2
`____________
`
`
`
`
`
`
`DECLARATION OF GEORGE EDWARDS
`IN SUPPORT OF PATENT OWNER’S RESPONSE
`
`
`
`
`
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 1
`
`

`

`
`
`TABLE OF CONTENTS
`
`i
`
`

`

`
`
`I, George Edwards, declare as follows:
`
`1.
`
`I am over the age of 21 years and am fully competent to make this
`
`Declaration. I make the following statements based on personal knowledge and, if
`
`called to testify to them, could and would do so.
`
`2.
`
`I have been retained on behalf of Sleep Number Corporation (“Sleep
`
`Number”). My fee is not contingent on the outcome of any matter or on any of the
`
`positions I have or will take in any matter. I have no financial interest in Sleep
`
`Number.
`
`1
`
`

`

`
`
`5.
`
`I attended graduate school at the University of Southern California,
`
`where I was a USC Viterbi School of Engineering Dean's Doctoral Fellow and
`
`Annenberg Graduate Fellow. I received an MS in computer science in 2006 and a
`
`PhD in computer science in 2010 from USC. My MS research focused on
`
`distributed, real-time, and embedded systems. My PhD research focused on the
`
`analysis of distributed systems and their architecture, with an emphasis on mobile
`
`applications and embedded systems. My research was funded by several government
`
`agencies, such as the Department of Defense and the NSA, and large companies,
`
`such as Bosch and InfoSys. I presented my work at numerous conferences and in
`
`academic journals, industry magazines, and other publications. In 2008, I received
`
`the USC Computer Science Department’s award for outstanding graduate student
`
`research.
`
`6.
`
`I am the founder of Quandary Peak Research, Inc., where I hold the
`
`titles of President and Computer Scientist. Quandary Peak Research is a software
`
`analysis company with twelve employees located in Los Angeles, Nashville, and
`
`Washington, D.C. Quandary Peak Research’s software analysis services fall into
`
`three areas: (1) audits and compliance in highly regulated industries; (2) technical
`
`due diligence in mergers and acquisitions; and (3) litigation and intellectual property.
`
`In all three areas, Quandary Peak analyzes software and computer systems to answer
`
`questions about the development, structure, behavior, and quality of those systems.
`
`2
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 4
`
`

`

`
`
`7.
`
`In my role as President, I manage the company’s business affairs. In
`
`my role as Computer Scientist, I perform software analysis on behalf of clients. I
`
`have analyzed a broad variety of complex, real-world software systems, including
`
`many embedded software systems, mobile devices, and web and Internet
`
`applications. I have reverse-engineered the hardware and software designs of dozens
`
`of distributed systems. I have conducted many investigations of the design and
`
`implementation of these systems with respect to specific patent claims and other
`
`intellectual property considerations, such as copyright infringement and trade secret
`
`theft.
`
`8.
`
`I was formerly employed as a Lecturer of Computer Science at the
`
`University of Southern California. In that capacity, I taught Requirements
`
`Engineering (CSCI 568), a graduate-level software engineering class, and Data
`
`Structures and Algorithms (CSCI 102), an undergraduate-level software design and
`
`programming class.
`
`9.
`
`I also formerly worked as a research scientist and software engineer at
`
`Blue Cell Software LLC, Intelligent Systems Technology, Inc., IBM, and The
`
`Boeing Company. During my time with Blue Cell, I built a simulation-based
`
`software design and modeling environment. While at IBM, I conducted research on
`
`next-generation mobile architectures, such as large-scale mobile device provisioning
`
`3
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 5
`
`

`

`
`
`systems. At Boeing, I helped to design software for collecting and processing sensor
`
`data and geo-location data from military vehicles.
`
`10.
`
`I have authored over thirty scholarly journal articles, magazine articles,
`
`conference papers, and book chapters on varied topics related to software
`
`engineering and distributed systems. I have co-authored many peer-reviewed papers
`
`specifically related to embedded systems and software control systems, such as
`
`“Integrating Publisher/Subscriber Services
`
`in Component Middleware
`
`for
`
`Distributed Real-time and Embedded Systems,”1 “Engineering Heterogeneous
`
`Robotics Systems: A Software Architecture-Based Approach,”2 and “A Middleware
`
`
`1 George Edwards, Douglas C. Schmidt, Aniruddha Gokhale, and Bala Natarajan.
`
`“Integrating Publisher/Subscriber Services in Component Middleware for
`
`Distributed Real-time and Embedded Systems.” Proceedings of the 42nd Annual
`
`ACM Southeast Conference (ACMSE), April 2004.
`
`2 Nenad Medvidovic, Hossein Tajalli, Joshua Garcia, Yuriy Brun, Ivo Krka, and
`
`George Edwards. “Engineering Heterogeneous Robotics Systems: A Software
`
`Architecture-Based Approach.” IEEE Computer, 2011.
`
`4
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 6
`
`

`

`
`
`Platform for Providing Mobile and Embedded Computing Instruction to Software
`
`Engineering Students,”3 among others.
`
`11.
`
`I have delivered numerous invited lectures, seminars, and technology
`
`demonstrations related to software design and analysis for university courses,
`
`research symposia, conferences, workshops, and industry events in the field of
`
`computer science. I have also served as a reviewer, committee member, or panelist
`
`for over a dozen computer science journals, magazines, and conferences.
`
`12.
`
`I am also a member of the Association for Computer Machinery (ACM)
`
`and the IEEE Computer Society.
`
`5
`
`

`

`
`
`and methods that are implemented and executed to a large degree in software that
`
`runs on a microprocessor. For example, the ‘154 and ‘747 Patents teach that the
`
`invention includes a “control device” and “control logic”:
`
`“The control device includes control logic that is capable of calculating
`
`a manifold pressure target based upon the desired pressure setpoint
`
`and a pressure adjustment factor, monitoring pressure within the pump
`
`manifold, adjusting pressure within the air chamber until the sensed
`
`manifold pressure is within an acceptable pressure target error range
`
`of the manifold pressure target, comparing an actual chamber pressure
`
`to the desired pressure setpoint to quantify an adjustment factor error,
`
`and calculating an updated pressure adjustment factor based upon the
`
`adjustment factor error.”4
`
`14. This disclosed “control logic” would naturally be implemented in
`
`software. The control logic is explicitly recited in the system claims of the ‘154 and
`
`‘747 Patents, and the disclosed capabilities of the control logic include the steps
`
`recited in the method claims of the ‘154 and ‘747 Patents.
`
`15. The ‘172 Patent also claims systems that include software and methods
`
`that are performed partly in software. For example, the ‘172 Patent teaches that the
`
`
`4 154 Patent at 2:50-60; 747 Patent at 2:40-50.
`
`6
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 8
`
`

`

`
`
`invention includes “a processor for providing commands to the improved valve
`
`enclosure assembly during an inflate/deflate cycle.”5 This processor would
`
`necessarily need to execute a software program to provide commands during an
`
`inflate/deflate cycle.
`
`16. Therefore, I am qualified to provide expert opinions regarding the
`
`aspects of the Patents-at-Issue that relate to software.
`
`17.
`
`I am also familiar with certain versions of source code produced by
`
`Petitioner American National Manufacturing, Inc. (“ANM”) and Real-Party-in-
`
`Interest Sizewise Rentals, LLC (“Sizewise”). I was retained by Sleep Number to
`
`conduct source code inspections in the underlying district court cases: Sleep Number
`
`Corporation v. American National Manufacturing,
`
`Inc., No. 5:18-cv-
`
`00357AB(SPx) (C.D. Cal. 2018) and Sleep Number Corporation v. Sizewise Rentals,
`
`LLC, No. 5:18-cv-00356AB(SPx) (C.D. Cal.) (“District Court Cases”). I was also
`
`previously retained by Select Comfort Corporation to conduct source code
`
`inspections in connection with Certain Air Mattress Systems, Components Thereof,
`
`and Methods of Using the Same, US ITC Inv. No. 337-TA-971.
`
`18. During my inspection of ANM’s source code in the District Court
`
`Cases, I analyzed what I identified as Version 1.8, Version 1.9, Version 1.92,
`
`
`5 172 Patent at 2:59-61.
`
`7
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 9
`
`

`

`
`
`Version 1.97, and Version 2.0 of ANM’s source code. My inspection included an
`
`analysis of the source code in comparison to the claims of the Patents-at-Issue.
`
`While I inspected and analyzed other versions of source code from ANM and
`
`Sizewise and compared them to the claims of the Patents-at-Issue, I do not discuss
`
`that analysis herein.
`
`19.
`
`It is my understanding that, in the District Court Cases, Sleep Number
`
`has accused various of ANM’s consumer and medical air bed products that use the
`
`source code identified above of infringing the Patents-at-Issue. It is my
`
`understanding that ANM petitioned the United States Patent and Trademark Office
`
`for IPR on each of the Patents-at-Issue. It is also my understanding that an IPR has
`
`been instituted on each patent.
`
`20.
`
`In my declaration, I rely on Sleep Number’s Amended Disclosure of
`
`Asserted Claims and Infringement Contentions against ANM, which is Exhibit
`
`2044; the ‘172 Patent claim chart against ANM, which is Exhibit 2046; the ‘154
`
`Patent claim chart against ANM, which is Exhibit 2048; and the ‘747 Patent claim
`
`chart against ANM, which is Exhibit 2049 (collectively herein “Contentions”). The
`
`Contentions contain reference to source code lines numbers and variable and method
`
`names that I helped compile, but I understand those references have been redacted
`
`and are not being provided to the Board at this time.
`
`8
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 10
`
`

`

`
`
`21.
`
`I incorporate the Contentions by reference. I do not provide any
`
`opinions on infringement in this declaration, as I understand infringement is not at
`
`issue in this IPR proceeding.
`
`22.
`
`I also reviewed the source code related to Sleep Number’s products that
`
`I understand Sleep Number located and provided for inspection to ANM’s counsel
`
`in the District Court Cases. I understand that Sleep Number has disclosed in its
`
`Contentions which air controllers sold with its consumer air mattress systems
`
`practice the claims of the Patents-at-Issue.
`
`23.
`
`I understand the source code provided for inspection by ANM and
`
`Sizewise in the District Court Cases is governed by protective orders under which
`
`the source code is marked “HIGHLY CONFIDENTIAL SOURCE CODE –
`
`OUTSIDE COUNSEL ONLY.” I take my obligations under the District Court
`
`Cases’ protective orders very seriously and have maintained the confidentiality of
`
`ANM’s and Sizewise’s source code.
`
`24.
`
`I have also reviewed a document produced by ANM in the District
`
`Court Cases with the Bates number ANMI00133414–22, which I understand is
`
`Exhibit 2052, and a document produced by Sleep Number in the District Court Cases
`
`with a Bates number SN_0021013–33, which I understand is Exhibit 2058.
`
`9
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 11
`
`

`

`
`
`10
`
`

`

`
`
`27. Static program analysis was an appropriate method to use in this matter
`
`because (1) it provides information highly relevant to the questions that I was asked,
`
`and (2) it only requires access to the source code, and does not require access to a
`
`running instance of the software. Static analysis was therefore an appropriate
`
`approach, given that the form of the software production and restrictions delineated
`
`in the Protective Order did not allow for deployment and execution of the software.
`
`28. Numerous static program analysis tools exist which analyze different
`
`aspects of the code and provide different perspectives on its structure and behavior.
`
`In my analysis of the ANM, Sizewise, and Sleep Number source code, I used
`
`Notepad++6, Eclipse7, and MPLAB X8 for viewing and printing the source code files.
`
`I used PowerGREP9 for searching the source code files. I used Beyond Compare10
`
`for comparing source code files and folders.
`
`
`6 Notepad++ 7 (version 7.7.1), https://notepad-plus-plus.org/.
`
`7 Eclipse IDE for C/C++ Developers,
`
`https://www.eclipse.org/downloads/packages/release/oxygen/2/eclipse-ide-cc-
`
`developers.
`
`8 MP LAB X IDE, https://www.microchip.com/mplab/mplab-x-ide.
`
`9 PowerGREP 5, https://www.powergrep.com/.
`
`10 BeyondCompare 4 (version 4.2.9), https://www.scootersoftware.com/.
`
`11
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 13
`
`

`

`
`
`12
`
`

`

`
`
`• The similarity between ANM’s and Sleep Number’s products and the
`
`release of the relevant ANM products after Sleep Number’s products are
`
`evidence that ANM copied the functionality claimed in the ‘172, ‘154, and
`
`‘747 Patents from the Sleep Number products.
`
`• ANM’s changes in moving from Version 1.8 to Version 1.9 of its source
`
`code and subsequent reversion of those changes in Version 1.97 and
`
`Version 2.0 demonstrate the technical value of the inventions disclosed in
`
`the ‘154 and ‘747 Patents and are evidence that these inventions
`
`substantially improved the software-based control of inflation and
`
`deflation in ANM’s products.
`
`13
`
`

`

`
`
`embedded within the computing hardware – which is itself part of a larger
`
`mechanical system – the software may be referred to as “firmware” and the ANM
`
`and Sleep Number products may also be characterized as “embedded systems.”
`
`31.
`
`In the ANM and Sleep Number products, the software embedded in the
`
`microcontroller implements control functions for inflation and deflation, including
`
`responding to user inputs to change the pressure in the bed, obtaining readings of the
`
`current pressure in the bed, sending commands to valves and pumps, and displaying
`
`messages on a user interface. Therefore, the software is a critical component of the
`
`ANM and Sleep Number products and understanding the design and function of
`
`these products requires an understanding of the software.
`
`32. By implementing inflation and deflation control functions in software,
`
`it is possible to program sophisticated behaviors into the system. For example, in
`
`certain ANM products, inflation and deflation control functions are partly regulated
`
`by parameters stored in the memory of the microcontroller. These parameters are
`
`updated using pressure readings that are taken both during and after inflation and
`
`deflation. By dynamically updating these parameters during operation, the software
`
`may change its inflation and deflation behavior based on usage conditions. For
`
`example, the ‘154 and ‘747 Patents explain, and I confirmed in a discussion with Dr.
`
`Abraham, that these parameters would change depending on the weight of the user.
`
`14
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 16
`
`

`

`
`
`In this way, the inflate/deflate behavior of the ANM products is customized for a
`
`user’s weight – a capability made possible through the use of software-based control.
`
`15
`
`

`

`
`
`software elements of the ‘172 Patent. This code also implements what ANM refers
`
`to in Ex. 2052 as a “Target pressure” controller, meaning that the user first sets a
`
`desired pressure, and then, without additional user input, the system adjusts the
`
`pressure to achieve that target pressure. Other than the Pegasus code, all the other
`
`Sleep Number code provided for inspection includes software elements of the ‘747
`
`and ‘154 Patents in addition to the ‘172 Patent. This other Sleep Number code
`
`implements an improved “Target pressure” controller by using an adjustment factor
`
`as taught in the ‘747 and ‘154 Patents.
`
` ‘172 Patent Against ANM
`
`35.
`
`It is my opinion based upon the inspections I have conducted that at
`
`least Version 1.8, Version 1.97, and Version 2.0 of ANM’s source code implement
`
`the software elements of claims 2, 6, 12, 16, 20, 22, and 24 of the ‘172 Patent, and
`
`at least Version 1.9 and Version 1.92 of ANM’s source code implement the software
`
`elements of claims 12 and 16 of the ‘172 Patent. Portions of these versions of source
`
`code demonstrate how to monitor pressure and provide commands during an
`
`inflate/deflate cycle as required by the claims of the ‘172 Patent. Table 1 indicates
`
`the specific claim elements of the ‘172 Patent that are implemented in ANM’s source
`
`code.
`
`36.
`
`It is also my opinion that the Contentions, in Ex. 2046 (on the ‘172
`
`Patent against ANM), accurately and specifically identify the source code files, lines,
`
`16
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 18
`
`

`

`
`
`functions, and variables in ANM’s source code that read on the claims of the ‘172
`
`Patent. I understand that reference in the Contentions to the source code line
`
`numbers and variable and method names have been redacted and are not being
`
`provided to the Board at this time. I can attest that the redacted text specifically
`
`identifies the parts of the source code that implement each claim element listed. I am
`
`willing and able to further explain how each version of source code operates if
`
`allowed to do so.
`
`37.
`
`I have reviewed the Declaration of John Abraham that I understand is
`
`being submitted by Sleep Number contemporaneously with this declaration. I
`
`understand that Dr. Abraham opines that ANM’s products read on the claimed
`
`mechanical-related elements of (i.e. structural components of) the ‘172 Patent. I rely
`
`upon that opinion and do not provide any opinions on mechanical-related elements
`
`myself.
`
`38. Therefore, based upon my review of the ‘172 Patent’s claims, the
`
`Declaration of John Abraham (and discussions with Dr. Abraham), and ANM’s
`
`source code, it is my opinion that ANM’s products that use Version 1.8, Version 1.9,
`
`Version 1.92, Version 1.97, and Version 2.0 of ANM’s source code can provide
`
`further evidence of practicing the claims of the ‘172 Patent that are identified in
`
`Sleep Number’s Contentions.
`
`17
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 19
`
`

`

`
`
`39.
`
`It is also my opinion based upon reviewing Sleep Number’s source code
`
`that ANM’s Version 1.8, Version 1.97, and Version 2.0 software is substantially
`
`similar to the inspected Sleep Number software that post-dates patent issuance with
`
`respect to the software-based control functions that are claimed in the ‘172 Patent.
`
`Further, it is my understanding that each of these versions of software were released
`
`after Sleep Number’s similar software had been put into products being offered for
`
`sale.
`
`I understand
`
`from
`
`reviewing
`
`the Contentions, SN_0021013–33,
`
`ANMI00133414–22, and Dr. Abraham’s Declaration that ANM’s air mattress
`
`systems and air controllers are similar to Sleep Number’s air mattress systems and
`
`air controllers in terms of the mechanical, i.e., structural, components.
`
`40. Therefore, it is my opinion that the similarity between ANM’s and
`
`Sleep Number’s software (established by my analysis), the similarity between
`
`ANM’s and Sleep Number’s mechanical components (established by Dr. Abraham),
`
`and the release of the relevant ANM products after Sleep Number’s products are
`
`evidence that ANM copied the functionality claimed in the ‘172 Patent from the
`
`Sleep Number products.
`
` ‘154 and ‘747 Patents Against ANM
`
`41.
`
`It is my opinion based upon the inspections I have conducted that at
`
`least Version 1.8, Version 1.97, and Version 2.0 of ANM’s source code read on the
`
`software-related limitations of claims 1-19 of the ‘154 Patent and claims 1-18 of the
`
`18
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 20
`
`

`

`
`
`‘747 Patent. The source code shows how the software receives a selection of a
`
`desired pressure setpoint, calculates a pressure target based upon the desired pressure
`
`setpoint and an adjustment factor, adjusts pressure within the air chamber of the air
`
`bed until a sensed pressure in a pump manifold is substantially equal to the pressure
`
`target, determines an actual pressure within the air chamber which is used to
`
`determine an adjustment factor error, and subsequently modifies the pressure
`
`adjustment factor based upon the adjustment factor error. Each version of the source
`
`code reads on other claimed elements, including but not limited to using an additive
`
`adjustment factor for inflation and a multiplicative adjustment factor for deflation,
`
`storing the modified/updated pressure adjustment factor, and using the modified
`
`pressure adjustment factor to calculate a modified/updated pressure target that is
`
`used in a subsequent pressure adjustment. Table 2 and Table 3 indicate the specific
`
`claim elements of the ‘154 and ‘747 Patent that are implemented in ANM’s source
`
`code.
`
`42.
`
`It is also my opinion that the Contentions, in Ex. 2048 (on the ‘154
`
`Patent against ANM) and Ex. 2049 (on the ‘747 Patent against ANM), accurately
`
`and specifically identify the source code files, lines, functions, and variables in
`
`Version 1.8, Version 1.97, and Version 2.0 of ANM’s source code that read on the
`
`claims of the ‘154 and ‘747 Patents. I understand that reference in the Contentions
`
`to the source code lines numbers and variable and method names have been redacted
`
`19
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 21
`
`

`

`
`
`and are not being provided to the Board at this time. I can attest that the redacted
`
`text specifically identifies the parts of each version of the source code that implement
`
`each claim limitation, and I would be willing and able to further explain how each
`
`version of source code operates, if allowed to do so.
`
`43. As stated above, I have reviewed the Declaration of John Abraham that
`
`is being submitted by Sleep Number contemporaneously with this declaration. I
`
`understand that Dr. Abraham opines that the mechanical structures of ANM’s
`
`products read on the claimed mechanical-related elements of (i.e. structural
`
`components of) the ‘154 and ‘747 Patents. I rely upon those opinions herein and do
`
`not provide any opinions on mechanical-related elements myself.
`
`44. Therefore, based upon my review of the ‘154 and ‘747 Patents’ claims,
`
`the Declaration of John Abraham (and discussions with Dr. Abraham), and ANM’s
`
`source code, it is my opinion that ANM’s products that use Version 1.8, Version
`
`1.97, and Version 2.0 of ANM’s source code each practice every element of the
`
`claims of the ‘154 and ‘747 Patents that are identified in Sleep Number’s
`
`Contentions.
`
`45.
`
` It is also my opinion based upon reviewing Sleep Number’s source
`
`code that ANM’s Version 1.8, Version 1.97, and Version 2.0 software is
`
`substantially similar to the inspected Sleep Number software that post-dates patent
`
`issuance with respect to the software-based control functions that are claimed in the
`
`20
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 22
`
`

`

`
`
`‘154 and ‘747 Patents, e.g., both Sleep Number’s inflate/deflate control software and
`
`ANM’s inflate/deflate control software practice each step of the method claims.
`
`Further, it is my understanding that each of these versions of ANM’s software was
`
`released after Sleep Number’s similar software had been put into products being
`
`offered for sale. I understand based upon reviewing the Contentions, SN_0021013–
`
`33, ANMI00133414–22, and Dr. Abraham’s Declaration that ANM’s air mattress
`
`systems are similar to Sleep Number’s air mattress systems and air controllers in
`
`terms of their mechanical, i.e., structural, components.
`
`46. Therefore, it is my opinion that the similarity between ANM’s and
`
`Sleep Number’s software (established by my analysis), the similarity between
`
`ANM’s and Sleep Number’s mechanical components (established by Dr. Abraham),
`
`and the release of the relevant ANM products after Sleep Number’s products are
`
`evidence that ANM copied the functionality claimed in the ‘154 and ‘747 Patents
`
`from the Sleep Number products.
`
`21
`
`

`

`
`
`version that was incorporated into the most recent product. (See ANMI00133422.)
`
`Although another page of this document refers to Version 1.5 – 1.8 (see
`
`ANMI00133421), the cover page of the document only refers to Version 1.8 (see
`
`ANMI00133414), which is the version I inspected.
`
`48.
`
`I have made the following conclusions looking at ANMI00133414-22:
`
`It appears ANM incorporated Version 1.8 into its products from the middle of 2014
`
`through the middle of 2016. As discussed above and further articulated in the
`
`Contentions, it is my opinion that Version 1.8 reads on the claims of the ‘154 and
`
`‘747 Patents and is substantially similar to the inspected Sleep Number software in
`
`terms of inflate/deflate control functionality. It appears ANM stopped using Version
`
`1.8 and instead incorporated Version 1.9 and Version 1.92 into its products from the
`
`middle of 2016 through the middle of 2017. It appears that ANM quickly abandoned
`
`use of Version 1.9 and used Version 1.92 for a very limited amount of time.
`
`49.
`
`In my opinion, both Versions 1.9 and 1.92 appear to be a step
`
`backwards from a design perspective. Versions 1.9 and 1.92 do not allow the system
`
`to achieve the same level of speed and accuracy as Version 1.8. These versions also
`
`appear to be substantially different than the inspected Sleep Number software in that
`
`they do not incorporate the software-based inflate/deflate control functions claimed
`
`in the ‘154 and ‘747 Patents. Based on my understanding of the changes that were
`
`made in the source code between Version 1.8 and Version 1.9, I believe that air
`
`22
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 24
`
`

`

`
`
`controllers using Version 1.9 of the software would exhibit an undesirable behavior:
`
`these products would more frequently fail to reach the desired pressure after the first
`
`attempt at inflation or deflation. This would require the system to repeatedly open
`
`and close a solenoid valve and turn an air pump on and off, which would increase
`
`the amount of time needed to reach a desired pressure and result in a suboptimal user
`
`experience.
`
`50.
`
`In the middle of 2017 or earlier, ANM stopped using Versions 1.9 and
`
`1.92 and incorporated new software into its products. I understand that Version 1.97
`
`may have been incorporated at some point, which may have been before mid-2017.
`
`However, at least by mid-2017, Version 2.0 was incorporated into ANM’s products.
`
`As discussed above and further articulated in the Contentions, it is my opinion that
`
`both Versions 1.97 and 2.0 read on the claims of the ‘154 and ‘747 Patents and are
`
`substantially similar to the inspected Sleep Number software regarding the
`
`inflate/deflate functionality claimed in the ‘154 and ‘747 Patents.
`
`51.
`
`Indeed, in my opinion, Versions 1.97 and 2.0 of ANM’s source code
`
`are materially the same as Version 1.8 of ANM’s source code. These three versions
`
`have the benefit of technology that allows for improved inflate/deflate control
`
`functionality and allows the system to achieve the user’s desired pressure setting
`
`with greater speed and accuracy than Version 1.9 and Version 1.92.
`
`23
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 25
`
`

`

`
`
`52. Thus, for approximately one year, ANM incorporated a version of
`
`software that is different from Sleep Number’s in that it does not implement the
`
`software-based inflate/deflate control functions claimed in the ‘154 and ‘747
`
`Patents, but then ANM went back to using source code that is basically identical to
`
`its earlier source code and which, like Sleep Number’s software, reads on the ‘154
`
`and ‘747 Patents.
`
`53.
`
`In my experience as a software engineer and developer, programmers
`
`try to improve their source code with each successive iteration or version.
`
`Accordingly, one would not revert to an earlier version of source code unless there
`
`was some benefit or value to doing so, or (equivalently) there was some significant
`
`problem or drawback with the later version. Therefore, I believe ANM must have
`
`found a benefit (or avoided a major shortcoming) by reverting to a prior version of
`
`source code that read on Sleep Number’s software-related patents rather than
`
`continuing to use a version that did not. Therefore, in my opinion ANM’s actions
`
`demonstrate the technical value of the inventions disclosed in the ‘154 and ‘747
`
`Patents and are evidence that these inventions substantially improved the software
`
`or solved a significant problem.
`
`24
`
`

`

`
`
`Issue. My analysis also found that those versions of ANM’s software are similar to
`
`certain versions of Sleep Number’s software with respect to the claimed software-
`
`based control functions. Those findings, coupled with evidence that the relevant
`
`ANM products were released after the Sleep Number products, is evidence that
`
`ANM copied the claimed software-based control functions from the Sleep Number
`
`products. Finally, ANM’s changes to its source code and later reversion of those
`
`changes is evidence of the technical value of the claimed software-based control
`
`functions.
`
`I declare under penalty of perjury that the foregoing is true and correct.
`
`
`
`Dated: ___________, 2019
`
`
`
`
`George Edwards
`
`
`
`
`25
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 27
`
`

`

`1
`
`
`GEORGE EDWARDS
`
`
`
`Computer Scientist
`
`Quandary Peak Research
`
`205 S Broadway, Ste 300
`
`Los Angeles, CA 90012
`
`Phone: 323.545.3933
`Email: george@quandarypeak.com
`
`
`
`
`
`
`
`
`EDUCATION
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`• Ph.D. in Computer Science, University of Southern California
`
`• M.S. in Computer Science, University of Southern California
`
`• B.S. in Computer Science with Minor in Mathematics, Vanderbilt University
`
`Aug 2010
`
`May 2006
`
`Aug 2003
`
`EMPLOYMENT
`
`June 2012 – Present
`
`• Computer Scientist
`Quandary Peak Research, Los Angeles, CA
`– Providing software analysis and expert witness testimony in software-related litigation, including
`patent and copyright infringement, theft of trade secrets, breach-of-contract, and other matters.
`– Analyzing software intellectual property and patent portfolios for validity and infringement in the
`context of licensing and brokering negotiations, startup investments, and M&A.
`– Investigating software failures to determine the root cause and help clients understand whether and
`how the failure could have been avoided.
`– Applying advanced analytic techniques based on calibrated parametric models for valuation of
`software products and estimation of software development costs.
`– Documenting the architecture of software systems to identify structural similarities and differences
`among competing products and services and deduce the origin of software designs and code.
`
`• Lecturer of Computer Science
`Computer Science Department, University of Southern California, Los Angeles, CA
`– Teaching a graduate-level software engineering course for M.S. and Ph.D. students and practicing
`software and aerospace engineers (CSCI 568).
`– Previously taught an undergraduate-level course on programming fundamentals, algorithms, and
`data structures (CSCI 102).
`
`Jan 2012 – May 2017
`
`• Co-Founder and CEO
`Blue Cell Software, Los Angeles, CA
`– Led and managed the development of a next-generation modeling and analysis platform for complex
`software-intensive systems.
`– Developed advanced model-based algorithms for reliability, efficiency, and risk analysis.
`– Conceived and implemented the business development strategy and produced and presented
`investment proposals.
`
`
`Aug 2010 – June 2012
`
`Sleep Number Corp.
`EXHIBIT 2029
`IPR2019-00500
`Page 28
`
`APPENDIX A
`
`

`

`2
`
`Feb 2011 – Sept 2011
`• Post-Doctoral Researcher
`Center for Systems and Software Engineering, University of Southern California, Los Angeles, CA
`– Conducted research on enterprise integration mechanisms to support net-centric operations.
`– Investigated methods, processes, and tools for decomposition and refinement of SoS capability needs
`into detai