`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`____________
`
`
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`IMMERSION CORPORATION,
`Patent Owner.
`________________
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`________________
`
`IMMERSION CORPORATION’S
`
`PATENT OWNER RESPONSE
`
`
`
`
`
`
`
`
`
`
`Mail Stop “PATENT BOARD”
`Patent Trial and Appeal Board
`U.S. Patent and Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`10180151
`
`
`
`
`
`
`
`TABLE OF CONTENTS
`
`Page
`
`
`I.
`
`II.
`
`PRELIMINARY STATEMENT .................................................................. 1
`
`THE INVENTION OF THE '356 PATENT ................................................ 3
`
`III. CLAIM CONSTRUCTION ......................................................................... 6
`
`IV. GROUND 1: CLAIMS 1-3, 5, 7, 9-13, 15, 17, 19-23, 25 AND 26
`ARE NOT OBVIOUS OVER ROSENBERG 737 AND
`ROSENBERG 281...................................................................................... 12
`
`A.
`
`Petitioner has provided no compelling reason for a
`POSITA to use a lookup table in the Rosenberg 737’s
`system ............................................................................................... 17
`B. When considering Rosenberg 737’s dynamic haptic
`applications along with the design factors, the evidence
`does not support the Petitioner's conclusion that a
`POSITA would have modified Rosenberg 737 with
`Rosenberg 281’s lookup table. ......................................................... 25
`Rosenberg 281’s lookup table is not the claimed lookup
`table—additional modification to that lookup table would
`be necessary. ..................................................................................... 30
`The testimony of Petitioner's expert should be given
`relatively little weight ....................................................................... 36
`
`C.
`
`D.
`
`V.
`
`CONCLUSION ........................................................................................... 37
`
`
`
`10180151
`
`
`
`- i -
`
`
`
`
`
`TABLE OF AUTHORITIES
`
` Page(s)
`
`Cases
`In re Kahn,
`441 F.3d 977 (Fed. Cir. 2006) ............................................................................ 21
`
`KSR Int’l Co. v. Teleflex Inc.,
`550 U.S. 398 (2007) ...................................................................................... 14, 21
`
`W.L. Gore & Assocs., Inc. v. Garlock, Inc.,
`721 F.2d 1540 (Fed. Cir. 1983) ...................................................................... 3, 12
`
`
`10180151
`
`
`
`- ii -
`
`
`
`
`
`
`
`EXHIBIT LIST
`
`Immersion Ex.
`2001
`Immersion Ex.
`2002
`Immersion Ex.
`2003
`Immersion Ex.
`2004
`Immersion Ex.
`2005
`Immersion Ex.
`2006
`Immersion Ex.
`2007
`
`Immersion Ex.
`2008
`Immersion Ex.
`2009
`
`Declaration of Nathan J. Delson, Ph.D., dated October 11,
`2016
`Assignment of U.S. Patent Application Serial No. 09/487,737
`to Immersion Corporation
`Assignment of U.S. Patent No. 8,773,356 to Immersion
`Corporation
`Microsoft Computer Dictionary (5th ed. 2002) (Excerpt)
`
`Declaration of Nathan J. Delson, Ph.D., dated May 31, 2017
`
`Deposition Transcript of Dr. Patrick Baudisch
`
`Roudat, Rau, Sterz, Plauth, Lopes & Baudisch, Gesture
`output: Eyes-free output using a force feedback touch surface,
`CHI 2013 Conference Paper
`ITC Claim Construction Order
`
`Curriculum Vitae of Nathan J. Delson, Ph.D.
`
`10180151
`
`
`
`- iii -
`
`
`
`I.
`
`PRELIMINARY STATEMENT
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`Patent Owner Immersion Corporation (“Immersion” or “Patent Owner”)
`
`submits this Response to the Board’s Decision – Institution of Inter Partes Review
`
`(Paper 7) (“Institution Decision”) as requested by Apple, Inc. ("Petitioner"),
`
`entered January 11, 2017, of United States Patent No. 8,773,356 (“the '356
`
`patent”).
`
`Petitioner argues that Rosenberg 281’s single disclosure of a lookup table
`
`would have been obvious to combine with Rosenberg 737’s system. However, the
`
`decision of whether to implement a lookup table into a particular system is based
`
`on a complex series of factors that Petitioner’s expert admits he did not address.
`
`For example, Petitioner’s expert Dr. Baudisch admits that there are numerous
`
`alternatives to a lookup table that can be used to retrieving haptic effects. Dr.
`
`Baudisch only considered one such alternative in his declaration. Furthermore, Dr.
`
`Baudisch admits that there are many factors that a POSITA would consider in
`
`determining which approach to implement in any particular system. Dr. Baudisch
`
`only considered one of these many factors in his declaration. In fact, in designing
`
`his own haptic feedback system, Dr. Baudisch chose an approach different than a
`
`lookup table. Ex. 2005 at ¶¶ 61-62.
`
`Petitioner cannot sustain its argument that implementation of the claimed
`
`- 1 -
`
`
`
`10180151
`
`
`
`
`lookup table would have been obvious. For instance, when the factors Petitioner’s
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`own expert admits a POSITA would have considered are taken into account, the
`
`claimed lookup table would be an unattractive design choice for Rosenberg 737’s
`
`system, because a lookup table is ill-suited for using haptic effects that are based
`
`on time-varying parameters such as velocity and frequency. Ex. 2005 at ¶¶ 55-60.
`
`Furthermore, even if a POSITA had been motivated to combine Rosenberg
`
`281’s lookup table with Rosenberg 737, such a combination would not arrive at the
`
`limitations of the challenged claims because the haptic effects retrieved from
`
`Rosenberg 281’s lookup table are based on the position of a user object (in degrees
`
`of freedom) rather than on the interaction between "an object contacting a touch-
`
`sensitive input device" (e.g., a user’s finger) and a graphical object, as required by
`
`the ’356 patent. See IPR2016-00807 at 17 (decision denying institution of a
`
`petition for review of the ‘356 patent, and explaining that the lookup table is based
`
`on “the position of a mouse or joystick”). Additional and substantial modification
`
`of Rosenberg 281’s lookup table would thus be necessary to satisfy the claims of
`
`the ‘356 patent. Petitioner provides no reason that a POSITA would have engaged
`
`in these additional modifications, which are taught by neither Rosenberg 281 nor
`
`Rosenberg 737. Instead, Petitioner’s argument is based solely on impermissible
`
`hindsight reconstruction of the challenged claims using the disclosure of the ‘356
`
`10180151
`
`
`- 2 -
`
`
`
`
`
`patent a roadmap to craft their arguments. W.L. Gore & Assocs., Inc. v. Garlock,
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`Inc., 721 F.2d 1540, 1553 (Fed. Cir. 1983) (improper to use “that which only the
`
`inventor taught . . . against its teacher.”).
`
`Finally, as the Board recognized in its Institution Decision, Petitioner’s
`
`arguments rely upon the credibility of the statements of its expert, Dr. Baudisch.
`
`Paper No. 7 at 24 (explaining that institution is proper "viewing the conflicting
`
`expert testimony . . . in the light most favorable to Petitioner"). But Dr. Baudisch’s
`
`analysis should be given relatively little weight, because it did not consider the
`
`very factors that Dr. Baudisch admitted a POSITA would have weighed in
`
`designing a haptic feedback system.
`
`II. THE INVENTION OF THE '356 PATENT
`
`The ’356 patent relates to providing tactile sensations to a user interacting
`
`(e.g., using his or her finger) with graphical objects displayed on a touchscreen in
`
`mobile electronic devices such as mobile phones and PDAs. The ’356 patent
`
`teaches, among other things, systems and methods in which the mobile electronic
`
`device displays on the touchscreen one or more graphical objects (such as, for
`
`example, menus, softkeys of a keypad, etc.). See, e.g., Ex. 1001 at 11:11-63
`
`(describing an embodiment in which a user touches and interacts with graphical
`
`objects on a display and a controller provides a corresponding tactile sensation).
`
`10180151
`
`
`- 3 -
`
`
`
`
`
`The user can contact the touch-sensitive touchscreen with an object (e.g., a finger)
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`and interact with the graphical object. The device generates an actuator signal for
`
`providing a tactile sensation to the user based at least in part on the user’s
`
`interaction with the graphical object and haptic effect data in a lookup table stored
`
`in a memory of the device. The lookup table, among other things, associates data
`
`related to interactions with haptic effect data. The tactile sensation, for example,
`
`provides a cue to the user interacting with a graphical object displayed on the
`
`touchscreen.
`
`The specification describes how the lookup table allows the system to
`
`associate tactile sensations with different interactions on the device. “In one
`
`embodiment, this information is in the form of associations among the detected
`
`input data, the functions of the electronic device or apparatus, and the tactile
`
`sensations.” Id. at 14:21-25. “The controller, using the data obtained from
`
`monitoring the input device, reads the table and obtains the associated function and
`
`tactile sensation information.” Id. at 14:33-35. Thus, the lookup table contains
`
`haptic effect data corresponding to various interactions between the object
`
`contacting the touchscreen (e.g., the user’s finger) and graphical objects displayed
`
`on the touchscreen. “The storage memory includes a table in which input signals
`
`are associated with various haptic feedback signals. This is explained more fully
`
`10180151
`
`
`- 4 -
`
`
`
`
`
`in relation to Figs. 9-10.” Id. at 7:678:3.
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`Figure 9, reproduced below, shows an embodiment of a lookup table.
`
`The lookup table facilitates associating different interactions with haptic
`
`effect data for different tactile sensations. Declaration of Nathan J. Delson, Ph.D.
`
`
`
`(“Delson Decl.”) (Ex. 2005) ¶¶ 22-25.
`
`Claim 1 recites:
`
`A method, comprising:
`outputting a display signal configured to display a graphical
`object on a touch-sensitive input device;
`
`10180151
`
`
`- 5 -
`
`
`
`
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`receiving a sensor signal from the touch-sensitive input device,
`the sensor signal indicating an object contacting the touch-
`sensitive input device;
`determining the interaction between the object contacting
`the touch-sensitive input device and the graphical object; and
`generating an actuator signal based at least in part on the
`interaction and haptic effect data in a lookup table. (emphasis
`added)
`
`The claim requires determining an interaction between an object and a
`
`graphical object and generating an actuator signal based at least in part on the
`
`interaction and haptic effect data in a lookup table.
`
`III. CLAIM CONSTRUCTION
`
`The claims require the limitation “generating an actuator signal based on at
`
`least in part on the interaction [between the object contacting the touch-sensitive
`
`input device and the graphical object] and haptic effect data in a lookup table.”
`
`The Board declined to construe this term. Decision at 9. Even under the broadest
`
`reasonable interpretation standard, the language of the claims themselves requires
`
`that the referenced “lookup table” contain an association between the interaction
`
`and haptic effect data. In other words, information about both “the interaction and
`
`haptic effect data” must be in the lookup table. For example, in the underlying ITC
`
`action, the Chief ALJ construed “lookup table” to mean “data structure in the form
`- 6 -
`
`
`10180151
`
`
`
`
`of a table containing associations between interactions and haptic effect data.” Ex.
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`2008 at 30 (noting that “the lookup table contains an association between the
`
`interaction and the haptic effect data, as claimed in the patent”). The interpretation
`
`adopted by the ITC is also the broadest reasonable construction under the standard
`
`for an inter partes review, because it is compelled by the claim language and the
`
`intrinsic record.
`
`First, the very wording of the claims compels an association between
`
`interaction and haptic effect data. Ex. 2005 ¶¶ 31-32. Claim 1 is recited below.
`
`A method, comprising:
`1.
`outputting a display signal configured to display a graphical
`object on a touch-sensitive input device;
`receiving a sensor signal from the touch-sensitive input device,
`the sensor signal indicating an object contacting the touch-sensitive
`input device;
`determining an interaction between the object contacting the
`touch-sensitive input device and the graphical object; and
`generating an actuator signal based at least in part on the
`interaction and haptic effect data in a lookup table. (emphasis
`added)
`
`The actuator signal is based on the interaction "and" haptic effect data, not
`
`based on the interaction "or" haptic effect data. Both types of data are used to
`
`10180151
`
`
`- 7 -
`
`
`
`
`
`generate the actuator signal consequentially there must be an association between
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`the interaction and haptic effect data.
`
`This construction is also supported by intrinsic evidence. Ex. 2005 ¶ 33.
`
`The ’356 patent describes the lookup table as a data structure located in memory.
`
`Ex. 1001 at 14:16-20 (“[T]he controller then accesses a memory device 54 in
`
`which is stored at least one database containing information necessary to produce
`
`the desired function in the electronic device and the predetermined tactile sensation
`
`in an input device, and accesses this information 55.”) (emphasis added). The ’356
`
`patent also explains that the lookup table contains associations. Id. at 14:23-25
`
`(“An exemplars [sic] group of associations is represented in tabular form in FIG.
`
`9.”) (emphasis added). For example, the specification teaches:
`
`[B]ased upon the information in the table associated with
`Pressure 1, the controller obtains the associated function
`information for selecting the number ‘2’, and information for
`distinct tactile Sensation 13. . . . The controller uses the
`information for distinct tactile Sensation 13 to produce
`Sensation 13 in an input device 56, by for example, causing an
`actuator to cause the input device to vibrate at a frequency
`associated with Sensation 13.
`
`Id. at 14:40-50 (emphasis added).
`
`10180151
`
`
`- 8 -
`
`
`
`
`
`Figure 9, which is an embodiment of the lookup table, reflects this
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`association between Pressure 1 and Sensation13, annotated below.
`
`
`
`In this example, the “PRESSURE DATA” column is associated with the
`
`haptic effect data contained in the “TACTILE SENSATION” column. Ex. 2005 at
`
`¶ 35. A POSITA would understand that the above example describes associating
`
`data relating to an interaction with haptic effect data (“tactile sensation”). Ex.
`
`2005 at ¶ 35.
`
`Because the actuator signal is controlled by the haptic effect data that is
`
`obtained from the lookup table, the lookup table must associate an “interaction”
`
`10180151
`
`
`- 9 -
`
`
`
`
`
`with haptic effect data if the resulting actuator signal is to be generated “based at
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`least in part on the interaction and haptic effect data in a lookup table.” Id.
`
`This claim scope was confirmed in the prosecution history of the ’356
`
`patent. Ex. 2005 ¶ 36. During prosecution, the applicant added to the last
`
`limitation of claim 1 “generating an actuator signal based at least in part on the
`
`interaction” the words “and haptic effect data in a lookup table” to overcome a
`
`rejection over the Rosenberg prior art. Ex. 1004, Page 43 (February 10, 2014
`
`Applicant Response at 2). In doing so, the applicant stated:
`
`Rosenberg may discuss outputting haptic effects based on user
`inputs (or graphical objects), but it does not discuss determining
`which specific haptic effect to output for given a (sic) user
`input (or graphical object) based on data in a lookup table."
`Id., Page 50 (February 10, 2014 Applicant Response, at 9)
`(emphasis added). In other words, the applicant distinguished
`the prior art because it did not disclose determining which
`haptic effect to output, based on a given interaction between the
`user and a graphical object, using a lookup table.
`
`Id. at 50.
`
`Petitioner incorrectly argues that “generating an actuator signal based at
`
`least in part on the interaction and haptic effect data in a lookup table” should be
`
`construed to mean “generating an actuator signal based at least in part on (1) the
`
`10180151
`
`
`- 10 -
`
`
`
`
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`interaction and (2) haptic effect data in a lookup table.” Pet. at 17. In other words,
`
`Petitioner contends “the only thing the claimed lookup table must include is ‘haptic
`
`effect data.’” Pet. at 18. Petitioner is incorrect.
`
`As explained above, in order for the generated actuator signal to be based at
`
`least in part on “the interaction” of the claim language, the ’356 patent requires that
`
`the lookup table contains an association between an interaction and haptic effect
`
`data. Ex. 2005 ¶ 38. Petitioner and its expert do not recognize that information
`
`about “the interaction” is necessary to identify the appropriate haptic effect data to
`
`output if the resulting actuator signal is to be based at least in part on the
`
`interaction. Pet. 18; Ex. 2005 at ¶ 38 (explaining that the lookup table must use
`
`data related to the interaction as input to obtain the resulting haptic effect data).
`
`By contending the claim can be satisfied with a lookup table containing no
`
`information about the interaction, Petitioner effectively’ reads out “the interaction”
`
`limitation.
`
`For the reasons stated above, when considered in context of the claims and
`
`the intrinsic record, the broadest reasonable construction of “generating an actuator
`
`signal based on at least in part on the interaction and haptic effect data in a lookup
`
`table” is that the claimed “lookup table” is a “data structure in the form of a table
`
`containing associations between interactions and haptic effect data.”
`
`10180151
`
`
`- 11 -
`
`
`
`
`
`IV. GROUND 1: CLAIMS 1-3, 5, 7, 9-13, 15, 17, 19-23, 25 AND 26 ARE
`NOT OBVIOUS OVER ROSENBERG 737 AND ROSENBERG 281.
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`Petitioner’s expert admits that Rosenberg 737 is not relied upon for any
`
`disclosure of a lookup table. Ex. 2006 at 31:25-32:6 ("the 737 does not disclose
`
`any lookup tables"). Instead, only the Rosenberg 281 reference supposedly
`
`provides the claimed lookup table. But Petitioner advances no compelling reason
`
`why a POSITA would use Rosenberg 281’s lookup table, modify it substantially so
`
`that it met the requirements of the ‘356 patent, and then incorporate it into
`
`Rosenberg 737’s system. Petitioner’s argument amounts to saying that because a
`
`lookup table arranged in one format is found in Rosenberg 281, a POSITA would
`
`have found it obvious to implement a lookup table in a different format Rosenberg
`
`737’s system. This is a clear case of hindsight reconstruction, which is
`
`impermissible. W.L. Gore & Assocs., Inc. v. Garlock, Inc., 721 F.2d 1540, 1553
`
`(Fed. Cir. 1983) (improper to use “that which only the inventor taught . . . against
`
`its teacher.”).
`
`In fact, the decision of whether to implement a lookup table into a particular
`
`system is based on a series of factors that Petitioner’s expert admits he did not
`
`address. For example, Petitioner’s expert Dr. Baudisch admits that there are
`
`numerous alternatives to retrieving haptic effects. Dr. Baudisch only considered
`
`10180151
`
`
`- 12 -
`
`
`
`
`
`one such alternative in his declaration. Furthermore, Dr. Baudisch admits that
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`there are many factors that a POSITA would consider in determining which
`
`approach to implement in any particular system. Dr. Baudisch only considered one
`
`of these many factors in his declaration. In fact, in designing his own haptic
`
`feedback system, Dr. Baudisch chose an approach different than a lookup table.
`
`Petitioner’s suggestion that implementation of a lookup table would have
`
`been obvious, in the absence of consideration of the factors Petitioner’s own expert
`
`admits a POSITA would have considered, rings hollow. Petitioner’s expert, Dr.
`
`Baudisch, agreed that a POSITA would not always use a lookup table. Ex. 2006 at
`
`19:19-20:1 ("it depends on a variety of factors"). He further testified that before
`
`deciding whether a lookup table was appropriate for any specific situation, a
`
`POSITA would have considered numerous requirements for a particular system
`
`such as "run-time behavior" or “speed,” “available memory on the system,”
`
`“maintenance” of the system, and suitability to the particular application at hand.
`
`Ex. 2006 at 20:18-21:5; 52:5-9; 54:3-18. Out of all of these factors, Petitioner
`
`considered only one (run-time behavior), and even in that instance the analysis was
`
`so conclusory and simplified that Dr. Baudisch admitted that it was not strictly
`
`true. See Pet. at 44 (contending that a "lookup table would be more efficient than
`
`other alternatives"); Ex. 2002 at ¶ 105 (same); Ex. 2006 at 33:4-17 (clarifying that
`
`10180151
`
`
`- 13 -
`
`
`
`
`
`this greater efficiency of a lookup table would only be true “[i]n some cases,” and
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`that a POSITA designing a system would have considered options other than a
`
`lookup table).
`
`Consideration of these factors is not a trivial point. In fact, Petitioner’s own
`
`expert declined to use a lookup table in a haptic feedback system he designed. Ex.
`
`2005 at ¶¶ 60-61. Thus, even if a POSITA may have had the skill necessary to
`
`implement a lookup table as taught by Rosenberg 281 in the Rosenberg 737’s
`
`system, Petitioner has identified no reason why a POSITA would have modified
`
`Rosenberg 737 in this manner. At most, Petitioner shows that a lookup table is one
`
`of many alternatives that may have been considered by a POSITA. Ex. 2005 at
`
`¶ 40. But simply showing that a particular feature is “a matter of design choice” is
`
`insufficient to show that a POSITA would have had reason to combine that
`
`implementation with that feature with the embodiment disclosed by another
`
`reference. See, e.g., IPR2015-2015-00613, Paper 11 at 15 (quoting KSR Int’l Co.
`
`v. Teleflex Inc., 550 U.S. 398, 418 (2007), and explaining that it is insufficient for
`
`obviousness to allege a particular combination of features would be “a matter of
`
`design choice”). Indeed, Rosenberg himself did not see fit to use a lookup table in
`
`the description of his Rosenberg 737 Application, despite the fact that he was
`
`aware of a lookup table (from Rosenberg 281).
`
`10180151
`
`
`- 14 -
`
`
`
`
`
`Furthermore, even if a POSITA had been motivated to combine Rosenberg
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`281’s lookup table with Rosenberg 737, such a combination would not arrive at the
`
`limitations of the challenged claims under either party's construction because the
`
`haptic effects retrieved from Rosenberg 281’s lookup table are based on the
`
`position of a user object (in degrees of freedom) rather than on the interaction
`
`between an object contacting a touch-sensitive input device and a graphical object,
`
`as required by the ’356 patent. See IPR2016-00807 at 17 (decision denying
`
`institution of a petition for review of the ‘356 patent, and explaining that the
`
`lookup table is based on “the position of a mouse or joystick”). Indeed, in
`
`Rosenberg 281, the single reference to a lookup table is contained in local memory
`
`27, which is local to the interface device 14 (e.g., joystick), as shown below from
`
`Figure 1 of Rosenberg 281.
`
`Local memory 27, such as RAM and/or ROM, is preferably coupled
`to microprocessor 26 in interface device 14 to store instructions for
`microprocessor 26 and store temporary and other 10 data. For
`example, force profiles can be stored in memory 27, such as a
`sequence of stored force values that can be output by the
`microprocessor, or a look-up table of force values to be output based
`on the current position of the user object.
`
`
`Rosenberg 281 at page 11, lines 9-12 (emphasis added).
`
`10180151
`
`
`- 15 -
`
`
`
`
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`
`
`The local memory 27 would have no information about any purported interaction
`
`with a graphical object located on display device 20 of the separate home computer
`
`system 12 because interface device 14 has no knowledge of the position of any
`
`graphical object on the display device or whether the user object is interacting with
`
`any graphical object. Ex. 2005 at ¶¶ 41-42. That determination occurs in the host
`
`computer 12. In Rosenberg 281’s local control loop (i.e., the embodiment that
`
`discusses the lookup table), the actuator signal would be generated when the user
`
`physically moves the user object (such as a joystick) in any direction, irrespective
`
`10180151
`
`
`- 16 -
`
`
`
`
`
`of whether there are graphical objects being displayed on the screen and
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`irrespective of whether the user is interacting with any of the graphical objects.
`
`Ex. 2005 at ¶ 42.
`
`Additional and substantial modification of Rosenberg 281’s lookup table
`
`would thus be necessary to satisfy the claims of the ‘356 patent under either party's
`
`proposed claim construction. Petitioner provides no reason that a POSITA would
`
`have engaged in these additional modifications, which are taught by neither
`
`Rosenberg 281 nor Rosenberg 737. Instead, Petitioner’s argument is based solely
`
`on impermissible hindsight reconstruction of the challenged claims using the
`
`disclosure of the ‘356 patent a roadmap to craft their arguments.
`
`A.
`
`Petitioner has provided no compelling reason for a POSITA to use
`a lookup table in the Rosenberg 737’s system
`
`Rosenberg 737 indisputably does not disclose a lookup table. See Ex. 2006
`
`at 31:25-32:6 ("the 737 does not disclose any lookup tables"). For this reason,
`
`Petitioner necessarily relies on a combination of Rosenberg 737 with Rosenberg
`
`281’s lookup table to supposedly render the challenged claims obvious.1 However,
`
`1 As explained below in Section IV.C, even if Rosenberg 281’s lookup table were
`
`combined with Rosenberg 737, a POSITA still would not arrive at the challenged
`
`claims.
`
`10180151
`
`
`- 17 -
`
`
`
`
`
`Petitioner’s arguments regarding motivation to combine are unpersuasive—by the
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`admission of Petitioner’s own expert, the Petition does not address the majority of
`
`factors a POSITA would have needed to address in determining whether
`
`implementation of a lookup table would have been desirable in a haptic feedback
`
`system. See Ex. 2006 at 20:18-21:5; 52:5-9; 54:3-18 (listing various factors a
`
`POSITA would have considered in determining how to implement haptic feedback
`
`in a system).
`
`At his deposition, Petitioner’s expert Dr. Baudisch admitted that there were
`
`several alternatives to using a lookup table. For example, one alternative is to “do
`
`everything in code somehow, not have a particular data structure.” Ex. 2006 at
`
`20:18-21. Another alternative is to “store haptic effects as a sequence of stored
`
`force values,” which Rosenberg 281 itself teaches is an alternative to a lookup
`
`table. Ex. 2006 at 34:19-35:3 (“I would assume that the person of ordinary skill in
`
`the art could do that.”); Ex. 1013 at 11:10-12 (“For example, force profiles can be
`
`stored in memory 27, such as a sequence of stored force values that can be output
`
`by the microprocessor, or a look-up table of force values to be output based on the
`
`current position of the user object.”). Another alternative, as set forth by Dr.
`
`Baudisch in his declaration, is “calculating a force value each time one was
`
`needed.” Ex. 1002 at ¶105. Calculating a force value when it is needed is
`
`10180151
`
`
`- 18 -
`
`
`
`
`
`explicitly taught as an alternative by Rosenberg 281. See Ex. 1013 at 11:12-15
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`(“In addition, a local clock 29 can . . . provide timing data . . . for example, to
`
`compute forces output by actuators 30 (e.g., forces dependent on calculated
`
`velocities or other time dependent factors).”); Ex. 2005 ¶ 45.
`
`With the exception of calculating a force value each time one is needed,
`
`Petitioner does not address any alternatives to using a lookup table or explain why
`
`a POSITA would find a lookup table more efficient. See Pet. at 44; Ex. 1002 at
`
`¶ 105. For instance, neither Petitioner nor its expert addresses the relative
`
`efficiency of a lookup table to not using a particular data structure, or simply
`
`storing the haptic effects as a sequence of stored force values. Id.; Ex. 2005 ¶ 46.
`
`Furthermore, the one alternative that Petitioner actually does address is done
`
`in a wholly conclusory fashion, without any of the analysis that Dr. Baudisch
`
`admits a POSITA would have needed to consider in determining whether a lookup
`
`table was an appropriate solution for a given application. Indeed, as Dr. Baudisch
`
`admits, Petitioner’s assertion that “a POSITA would have understood that
`
`obtaining force values from a lookup table would be more efficient than other
`
`alternatives, such as calculating a force value each time one was needed” (Pet. at
`
`44) is simply not accurate. As Dr. Baudisch explained at his deposition, whether
`
`using a lookup table would be suitable for any particular application “depends on a
`
`10180151
`
`
`- 19 -
`
`
`
`
`
`variety of factors.” Ex. 2006 at 19:19-20:1. The factors to be considered would
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`include “run-time behavior” (or “speed”), “maintenance,” and various other design
`
`requirements such as “available memory on the system.” Ex. 2006 at 20:22-21:5;
`
`54:3-18. A POSITA would also consider “the application and the specific means
`
`available” to “determine how you would go about storing and quantizing the
`
`effects.” Ex. 2006 at 55:13-17.
`
`Dr. Baudisch engaged in none of this analysis in his declaration. Ex. 2005
`
`¶¶ 47-48. He did not address the memory footprint of a lookup table or compare it
`
`to any potential efficiency tradeoffs. Ex. 2006 at 33:18-22. He did not address
`
`additional maintenance requirements and complications that may be associated
`
`with adding a lookup table data structure to a system. See Ex. 2005 at ¶ 48. He
`
`did not compare tradeoffs between available memory on a system and run-time
`
`behavior. Ex. 2006 at 33:18-22 (admitting “memory footprint” of various
`
`implementations not addressed in Dr. Baudisch’s declaration). Nor did Dr.
`
`Baudisch consider “the application” of Rosenberg 737 in determining whether
`
`combining Rosenberg 737 with Rosenberg 281 would have been a desirable design
`
`choice. And even regarding run-time behavior, Dr. Baudisch admitted that this
`
`factor could, in some cases, lead a POSITA to choose options other than a lookup
`
`table. Ex. 2006 at 33:4-17 (clarifying that his opinion in his declaration that “a
`
`10180151
`
`
`- 20 -
`
`
`
`
`
`lookup table would have been more efficient than other alternatives” was only true
`
`Case IPR2016-01381
`Patent No. 8,773,356
`
`
`in certain instances, and that a POSITA “might consider other options as well”). In
`
`fact, no evidence or explanation is provided to verify Dr. Baudisch’s conclusory
`
`assertion that a lookup table would supposedly be more efficient than other
`
`alternatives.
`
`Petitioner’s argument is not sufficient to establish obviousness. In order to
`
`support an obviousness finding, there must be a clear and explicit articulation of
`
`reasons why the claimed invention would have been obvious. KSR International
`
`Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007). A “rejection on obviousness cannot
`
`be sustained with mere conclusory statements; instead, there must be some
`
`articulated reasoning with some rational underpinning to support the legal
`
`conclusion of obviousness. In re Kahn, 441 F.3d 977, 988 (Fed. Cir. 2006). See
`
`KSR International Co. v. Teleflex Inc., 550 U.S. 398, 418 (2007) (quoting Federal
`
`Circuit statement with approval).
`
`The necessary, articulated with some rational underpinning, is simply
`
`lacking from the Petition and corresponding expert declaration. For example,
`
`Petitioner does not even address the two alternatives to a lookup t