`
`Filed on behalf of Marvell Semiconductor, Inc.
`By:
`Lori A. Gordon
`
`Sterne, Kessler, Goldstein & Fox PLLC
`
`1100 New York Avenue, NW
`
`
`Washington, D.C.
`
`
`Tel: (202) 371-2600
`
`
`Fax: (202) 371-2540
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`
`
`
`
`
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`
`
`
`
`
`
`
`
`
`
`PETITION FOR INTER PARTES REVIEW
`OF U.S. PATENT NO. 6,108,388
`
`
`
`
`
`
`
`Mail Stop PATENT BOARD
`Patent Trial and Appeal Board
`U.S. Patent & Trademark Office
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`TABLE OF CONTENTS
`
`B.
`
`C.
`D.
`
`- i -
`
`
`I.
`II.
`III.
`
`V.
`
`Mandatory notices (37 C.F.R. §42.8(a)(1)). ...................................................... 2
`Grounds for standing (37 C.F.R. §42.104(a)). .................................................. 3
`Identification of challenge (37 C.F.R. §42.104(b)). .......................................... 3
`A.
`Citation of prior art. ................................................................................. 3
`B.
`Statutory grounds for the challenge. ....................................................... 4
`IV. The ’388 patent. ................................................................................................. 4
`A.
`The ’388 patent claims the application of the well-known Turbo
`Principle to well-known components of a communications
`receiver. ................................................................................................... 4
`Technical Overview of the ’388 patent. .................................................. 6
`1. Prior art receivers. ............................................................................ 6
`2. The claimed receiver of the ’388 patent. ......................................... 8
`Level of ordinary skill in the art. ...........................................................14
`Claim construction. ...............................................................................15
`1. The Board should construe “weighted [values/symbols]” as “non-
`binary [values/symbols].” .............................................................. 16
`2. The Board should construe the phrase “or i greater than 1” as “for
`i greater than 1.” ........................................................................... 17
`3. Means-plus-function terms. ........................................................... 18
`The prior art. .................................................................................................... 20
`A. Overview of Zhou. ................................................................................20
`B.
`Overview of Berrou. ..............................................................................23
`VI. Ground 1: The combination of Zhou and Berrou renders claims 1–4 and
`7–9 obvious. ..................................................................................................... 26
`A.
`The combination renders independent claim 9 obvious. .......................27
`1. Claim 9 overview. .......................................................................... 27
`2. Combination of Zhou and Berrou.................................................. 30
`3. The combination discloses the preamble. ...................................... 36
`4. The combination discloses the “supplying” step. .......................... 37
`
`

`

`B.
`
`C.
`
`D.
`E.
`F.
`
`G.
`
`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`5. The combination discloses the “performing” step. ....................... 37
`6. The combination discloses the “correcting” step. ......................... 38
`7. The combination discloses the “decoding” step. ........................... 42
`8. The combination discloses the “computing” step. ........................ 45
`9. The combination discloses the “delivering” step of claim 9. ........ 47
`The combination of Zhou and Berrou renders claim 1 obvious. ..........49
`1. The combination discloses the preamble. ...................................... 50
`2. The combination discloses the claimed modules. ......................... 50
`3. The combination discloses the ISI correction means. ................... 52
`4. The combination discloses the decoding means............................ 54
`5. The combination discloses the correction information computation
`means. ............................................................................................ 55
`The combination renders claim 2 obvious. ...........................................58
`1. The combination discloses the recited inputs and outputs. ........... 62
`2. The combination discloses the symbol input/output connections. 65
`3. The combination discloses the correction input/output connections.67
`4. The combination discloses the decoded symbol output
`connections. ................................................................................... 69
`The combination renders claim 3 obvious. ...........................................71
`The combination renders claim 4 obvious. ...........................................73
`The combination renders claim 7 obvious. ...........................................75
`1. The combination discloses reception of interleaved signals. ........ 78
`2. The combination discloses de-interleaving means. ....................... 79
`3. The combination discloses interleaving means. ............................ 80
`The combination renders independent claim 8 obvious. .......................81
`1. The combination discloses the preamble. ...................................... 81
`2. The combination discloses the recited inputs and outputs. ........... 82
`3. The combination discloses the ISI correction means. ................... 83
`4. The combination discloses the decoding means............................ 84
`5. The combination discloses the correction information computation
`means. ............................................................................................ 84
`
`- ii -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`VII. Conclusion. ...................................................................................................... 85
`
`- iii -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`TABLE OF AUTHORITIES
`
`
`Cases
`
`Novo Indus., L.P. v. Micro Molds Corp.,
`350 F.3d 1348 (Fed. Cir. 2003) .............................................................................. 17
`
`
`Phillips v. AWH Corp.,
`415 F.3d 1303 (Fed. Cir. 2005) .............................................................................. 15
`
`
`Statutes
`
`35 U.S.C. §102(a) ........................................................................................................ 4
`
`35 U.S.C. §102(b) ........................................................................................................ 4
`
`35 U.S.C. §112 ..................................................................................................... 15, 18
`
`35 U.S.C. §154(a)(2) .................................................................................................. 15
`
`Other Authorities
`
`MPEP §608.01(m) ............................................................................................... 18, 82
`
`Regulations
`
`37 C.F.R. §42.104(a) .................................................................................................... 3
`
`37 C.F.R. §42.104(b) ................................................................................................... 3
`
`37 C.F.R. §42.8(a)(1) ................................................................................................... 2
`
`37 C.F.R. §42.8(b)(3) ................................................................................................... 2
`
`
`
`
`- iv -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`Ex. No.
`1001
`
`1002
`1003
`
`1004
`1005
`
`1006
`1007
`1008
`
`1009
`
`1010
`
`1011
`
`1012
`
`1013
`
`1014
`
`1015
`
`1016
`
`1017
`
`EXHIBIT LIST
`
`Description
`U.S. Patent 6,108,388 to Douillard et al. with certificate of correction
`on claim 9
`File History for U.S. Patent 6,108,388
`Declaration of Dr. Zixiang Xiong in Support of Petition for Inter
`Partes Review of U.S. Patent No. 6,108,388
`Curriculum Vitae of Dr. Zixiang Xiong
`Zhou et al., “Decision-Feedback Equalization of Time-Dispersive
`Channels with Coded Modulation,” IEEE Trans. on Commun., vol.
`38, issue 1 (1990) (“Zhou”)
`French Patent No. 2675971B1 to Berrou
`Certified English Translation of Exhibit 1006 (“Berrou”)
`Hagenauer et al., “Iterative (‘Turbo’) Decoding of Systematic
`Convolutional Codes with the MAP and SOVA Algorithms,” ITG
`Fachbericht 130 (1994) (“Hagenauer”)
`Berrou et al. “Near Shannon Limit Error-Correcting Coding and
`Decoding: Turbo Codes,” IEEE Int’l Conf. on Commun. (1993)
`(“Berrou 1993”)
`Certification of English-Language Translation of French Patent No.
`2675971B1 to Berrou (Exhibit 1007)
`French Patent Application Publication No. 2675968A1 to Berrou et
`al. (“Berrou FR’68”)
`Koetter et al., “Turbo Equalization,” IEEE Signal Processing
`Magazine (2004)
`Laot et al., “Turbo Equalization: Adaptive Equalization and Channel
`Decoding Jointly Optimized,” IEEE J. Selected Areas in Commun.
`(2001)
`Complaint filed April 5, 2016 in Spectra Licensing Group, LLC v.
`Marvell Semiconductor, Inc. et al., Civil Case No. 3:16-cv-00817-
`DMS-MDD, SDCA
`Lin et al., Error Control Coding Fundamentals and Applications, 2d
`Ed. (2004) (excerpts) (“Lin”)
`Douillard et al., “Iterative Correction of Intersymbol Interference:
`Turbo-Equalization,” European Trans. on Telecomm., vol. 6, no. 5,
`(Sept.–Oct. 1995)
`Tuchler et al., “Turbo Equalization: An Overview,” IEEE Trans. on
`Info. Theory, vol. 57, no. 2 (2011)
`
`- v -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`Ex. No.
`1018
`
`1019
`1020
`
`1021
`1022
`
`1023
`
`1024
`
`1025
`1026
`
`1027
`
`1028
`
`1029
`
`1030
`
`1031
`1032
`
`1033
`
`1034
`
`Description
`Sklar, Digital Communications: Fundamentals and Applications,
`(1988) (excerpts) (“Sklar”)
`U.S. Patent 5,446,747 to Berrou (“Berrou ’747”)
`Hagenauer, “The Turbo Principle: Tutorial Introduction and State of
`the Art,” International Symposium on Turbo Codes, (1997)
`(“Hagenauer 1997”)
`(Not Used)
`Parsons, The Mobile Radio Propagation Channel (1992) (excerpts)
`(“Parsons”)
`Proakis, Digital Communications, 2d ed. (1989) (excerpts)
`(“Proakis”)
`Forney, “The Viterbi Algorithm,” IEEE Proceedings, vol. 61, no. 3,
`(1973) (“Forney”)
`U.S. Patent 5,406,570 to Berrou et al. (“Berrou ’570”)
`Hagenauer et al., “A Viterbi Algorithm with Soft-Decision Outputs
`and its Applications,” IEEE Global Telecomm. Conf. (1989)
`(“Hagenauer 1989”)
`Gersho et al., “Adaptive Cancellation of Intersymbol Interference for
`Data Transmission,” The Bell System Technical Journal, vol. 60, no.
`11 (Nov. 1981) (“Gersho”)
`Hagenauer et al., “Decoding ‘Turbo-Codes’ with the Soft Output
`Viterbi Algorithm,” IEEE International Symposium on Information
`Theory (1994) (“Hagenauer 1994”)
`Proakis, “Adaptive Non-Linear Filtering Techniques for Data
`Transmission,” IEEE Symposium on Adaptive Processes, Decision,
`and Control (1970) (“Proakis II”)
`Nowlan et al., “A Soft Decision-Directed LMS Algorithm for Blind
`Equalization,” IEEE Trans. on Commun., vol. 41, no. 2 (1993)
`(“Nowlan”)
`Declaration of Steven W. Peters, Ph.D.
`IEEE Xplore Database Entry for IEEE Trans. on Commun., vol. 38,
`issue 1 (1990)
`Goldsmith, “Design and Performance of High-Speed Communication
`Systems over Time-Varying Radio Channels,” University of
`California at Berkeley, Ph.D. Dissertation (1994) (citing Zhou as
`[61])
`Gozzo, “Recursive Least-Squares Sequence Estimation,” IBM J. Res.
`Develop., vol. 38, no. 2 (Mar. 1994) (citing Zhou as [21])
`
`- vi -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`Ex. No.
`1035
`
`1036
`
`1037
`
`1038
`
`1039
`
`Description
`Proakis, “Adaptive Equalization for TDMA Digital Mobile Radio,”
`IEEE Trans. on Vehicular Technology, vol. 40, no. 2 (May 1991)
`(citing Zhou as [12])
`Simmons, “Alternative Trellis Decoding for Coded QAM in the
`Presence of ISI,” IEEE Trans. on Commun., vol. 42, no. 2/3/4,
`(1994) (citing Zhou as [2])
`Copyright Registration Record, IEEE Trans. on Commun., vol. 38,
`Library of Congress.
`IEEE Trans. on Commun., vol. 38, no. 1 (Jan. 1990) Library of
`Congress
`Taylor, “The Estimate Feedback Equalizer: A Suboptimum
`Nonlinear Receiver,” IEEE Trans. on Commun., vol. 21, no. 9 (Sept.
`1973) (“Taylor”)
`
`- vii -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`Marvell Semiconductor, Inc. (“Marvell”) petitions for inter partes review of
`
`claims 1–4 and 7–9 of United States Patent No. 6,108,388 to Douillard et al. (the
`
`’388 patent). The ’388 patent discloses an iterative method for processing received
`
`signals in a digital receiver. Each iteration performs two steps, equalization and
`
`decoding. The method also calculates “correction information” that is passed
`
`between iterations to improve performance at each iteration. The ’388 patent
`
`acknowledges that equalization followed by decoding was known in the prior art.
`
`Zhou, an IEEE journal article, disclosed a two-iteration method for equalization and
`
`decoding five years before the ’388 patent.
`
`The challenged claims do not recite any details about the equalization or
`
`decoding steps. Rather, the claims recite how those steps relate to each other and
`
`what information they input and output. Those relationships, however, are borrowed
`
`directly from the well-known turbo codes disclosed by Berrou four years before
`
`the’388 patent. Acknowledging that origin, the inventors used the phrase “turbo
`
`equalization” in papers to describe the application of iterative decoding to
`
`equalization and decoding. Scholars have since noted that turbo equalization
`
`resulted from the straightforward application of turbo codes to the related problem
`
`of equalization.
`
`Marvell demonstrates below that a reasonable likelihood exists that the
`
`challenged claims are unpatentable over Zhou and Berrou.
`
`- 1 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`I. Mandatory notices (37 C.F.R. §42.8(a)(1)).
`REAL PARTY IN INTEREST: The real party-in-interest of Petitioner is Marvell
`
`Semiconductor, Inc.
`
`RELATED MATTERS: The ’388 patent is the subject of the following civil
`
`actions:
`
`Spectra Licensing Group, LLC v. Marvell Semiconductor, Inc. et al, Civil
`
`Case No. 3:16-cv-00817-DMS-MDD filed April 5, 2016 in the Southern District of
`
`California; Spectra Licensing Group, LLC v. LSI Corporation et al., Civil Action
`
`No. 3:15-cv-00899-DMS-MDD filed April 14, 2016 in the Southern District of
`
`California; Spectra Licensing Group, LLC v. LSI Corporation et al., Civil Action
`
`No. 3:16-cv-06095-RS filed October 21, 2016 in the Northern District of California;
`
`and Spectra Licensing Group, LLC v. Marvell Semiconductor, Inc. et al., Civil
`
`Action No. 3:16-cv-06093-RS filed October 21, 2016 in the Northern District of
`
`California.
`
`No other matters related to the ’388 patent are known to Marvell.
`
`LEAD AND BACKUP COUNSEL: Pursuant to 37 C.F.R. §42.8(b)(3) and
`
`42.10(a), Marvell appoints Lori A. Gordon (Reg. No. 50,633) as its lead counsel
`
`and Robert E. Sokohl (Reg. No. 36,013) and Steven W. Peters (Reg. No. 73,193)
`
`as its back-up counsel; all at the address: STERNE, KESSLER, GOLDSTEIN & FOX,
`
`- 2 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`1100 New York Avenue, N.W., Washington, D.C., 20005, phone number (202) 371-
`
`2600 and facsimile (202) 371-2540.
`
`SERVICE INFORMATION: Marvell consents to electronic service by email at
`
`the email addresses: lgordon-PTAB@skgf.com, rsokohl-PTAB@skgf.com, and
`
`speters-PTAB@skgf.com.
`
`II. Grounds for standing (37 C.F.R. §42.104(a)).
`The undersigned and Marvell certify that the ʼ388 patent is available for inter
`
`partes review. Marvell certifies that it is not barred or estopped from requesting this
`
`inter partes review on the grounds identified herein. The assignee of the ’388 patent,
`
`Spectra Licensing, filed a complaint against Marvell alleging infringement of the
`
`’388 patent on April 5, 2016. (Ex. 1014.) The present petition is being filed within
`
`one year of the April 7, 2016 service on Marvell.
`
`III.
`
`Identification of challenge (37 C.F.R. §42.104(b)).
`A. Citation of prior art.
`The ’388 patent is the U.S. national stage of international application
`
`PCT/FR96/00197, filed on February 6, 1996. The ’388 patent further claims priority
`
`to a French application, filed on February 7, 1995.1 In support of the grounds of
`
`unpatentability cited above, Marvell cites the following prior art references:
`
`
`1 Marvell does not acquiesce that the ’388 patent is entitled to benefit of the
`
`1995 French application.
`
`- 3 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`Decision-Feedback Equalization of Time-Dispersive Channels with
`
`Coded Modulation by Zhou et al., (Ex. 1005), is prior art under at least 35 U.S.C.
`
`§§102(a) and 102(b) because it published in January 1990, more than five years
`
`before the ’388 patent’s U.S. filing date. (See Exs. 1031–1037.)
`
`French Patent No. 2675971B1 to Berrou, provided in the original French as
`
`Ex. 1006 and translated into English as Ex. 1007, is prior art under at least 35 U.S.C.
`
`§§102(a) and 102(b) because it issued on August 6, 1993, more than one year before
`
`the ’388 patent’s U.S. filing date. (Ex. 1007, Berrou, Face, (45).)
`
`Statutory grounds for the challenge.
`
`B.
`Marvell requests review of claims 1–4, 7, 8, and 9 (as corrected) on the
`
`following grounds:
`
`References
`
`Basis Claims Challenged
`
`Zhou and Berrou
`
`§103
`
`1–4, 7–9
`
`
`IV. The ’388 patent.
`A. The ’388 patent claims the application of the well-known Turbo
`Principle to well-known components of a communications receiver.
`
`In 1991, one of the co-inventors of the ’388 patent, Claude Berrou, filed a
`
`French application on a technology he would later call “turbo codes.” (Xiong Decl.,
`
`¶24.) The primary innovation of turbo codes was decoding of a signal using multiple
`
`iterations, each iteration performing two decoding steps that output “extrinsic
`
`- 4 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`information.” This “extrinsic information” allowed each iteration to improve upon
`
`the previous one. (Id., ¶26.) Researchers subsequently termed this architecture the
`
`Turbo Principle. (Id.)
`
`In 1995, four years after disclosing the Turbo Principle, Berrou and some co-
`
`workers filed a second French application, applying the Turbo Principle, that would
`
`eventually result in the ’388 patent. (Id., ¶31.) The primary difference between turbo
`
`codes in the original Berrou patent and the system disclosed in the ’388 patent is
`
`that the iterations in the ’388 patent perform equalization and decoding rather than
`
`performing a first decoding and a second decoding. (Id., ¶44.) A subset of the
`
`inventors, along with other members of Berrou’s team, published a series of articles
`
`relating to the iteration over equalization and decoding, and the authors termed this
`
`technology “turbo equalization.” (Id., ¶31.)
`
`The inventors of the ’388 patent, however, were not the first to disclose
`
`iterative processing of equalization and decoding. Zhou disclosed a two-iteration
`
`equalization and decoding method as early as January 1990, even before Berrou’s
`
`1991 disclosure of turbo codes. (Xiong Decl., ¶52.) Zhou did not disclose the
`
`specific extrinsic information enabling the Turbo Principle. (Id.) However, Berrou’s
`
`existing and well-known turbo codes used extrinsic information in a manner that
`
`was readily applicable to Zhou’s architecture. (Id.)
`
`- 5 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`Marvell will show that the ’388 patent claims the obvious application of the
`
`Turbo Principle to well-known steps of receiver processing. Skilled artisans have
`
`noted this fact, describing the inventors’ work as “a natural extension” of the
`
`development of turbo codes that “leveraged the ideas of the turbo decoding
`
`algorithm to the related problem of equalization and decoding.” (Ex. 1012, Abstract,
`
`p. 68, col. 1, ¶1.) A paper co-authored by a co-inventor of the ’388 patent admits
`
`that the principle of the supposed invention was “borrowed from turbo-codes.” (Ex.
`
`1013, p. 1744, col. 2, ¶5.)
`
`B.
`
`Technical Overview of the ’388 patent.
`1. Prior art receivers.
`Error-correcting codes add redundancy to a transmitted signal to improve a
`
`receiver’s ability to detect and correct errors in the signal after it has traversed the
`
`channel. (Xiong Decl., ¶25.) Figure 1 of the ’388 patent (below) depicts a prior art
`
`receiver implementing a convolutive error correction decoding operation following
`
`equalization.
`
`In addition to a demodulator (not shown), the prior art receiver includes
`
`equalizer 19, de-interleaver 110, and convolutive decoder 111. Equalizer 19 acts as
`
`a means to correct inter-symbol interference (ISI). ISI was a well-known
`
`phenomenon prior to the ’388 patent. (Id., ¶¶29–30.) ISI results when echoes of a
`
`prior transmission (e.g., reflections of a transmission off objects near the transmitter
`
`- 6 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`or receiver) interfere with later transmissions. (Id., ¶29.) A receiver must correct for
`
`ISI when it exists using a component called an equalizer. (Id., ¶30.) The ’388 patent
`
`explains that in the prior art receiver, the received signal is “generally processed by
`
`an equalizer…given the task of eliminating the inter-symbol interference introduced
`
`by the channel.” (’388 patent, 1:21–25.)
`
`
`
`The output of the equalizer is “then de-interleaved if necessary and decoded
`
`before being given to the addressee.” (’388 patent, 1:25–27.) Demodulation, de-
`
`interleaving, and decoding perform reciprocal operations to their transmission-side
`
`counterparts. (Id., 8:32–39; Xiong Decl., ¶25.) Specifically, prior to transmission,
`
`the error-correction encoder of the transmitter adds redundancy to the binary data
`
`before interleaving (re-arranging) and modulating the data for transmission. (’388
`
`patent, 7:11–50.) Turbo Codes are an example of an error-correction encoding
`
`technique known prior to the ’388 patent. (Id., 2:33–40.)
`
`Decoder 111 complements the transmitter’s encoder by reversing the applied
`
`encoding, whereas the de-interleaver complements the transmitter’s interleaver by
`
`- 7 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`reversing the rearrangement of bits. In addition, decoder 111 corrects errors in the
`
`equalized signal (i.e., the output of the equalizer) that may have occurred due to
`
`noise, interference, or other distortion. (Xiong Decl., ¶25.) Figure 1 focuses on a
`
`convolutive decoder, implying that the encoder is a convolutional encoder. (Id.,
`
`¶41.) The ’388 patent specifies that the convolutive decoder may “implement[] an
`
`operation of maximum likelihood decoding such as the Viterbi algorithm.” (’388
`
`patent, 8:40–43.) The Viterbi algorithm can also be used in other applications
`
`separate from a convolutive decoder, such as sequence detection for ISI mitigation.
`
`(Id., 2:9–11.) The output of the convolutive decoder is provided to the recipient
`
`device (addressee).
`
`2. The claimed receiver of the ’388 patent.
`The ’388 patent makes a known modification to the conventional receiver
`
`structure by iterating between equalization and decoding at least twice. In a first
`
`iteration, the receiver equalizes and decodes similar to the conventional receiver
`
`discussed in the previous section. Rather than passing the decoded signal to the
`
`recipient device, the receiver uses it in a second round, or iteration, of equalization
`
`and decoding. (Xiong Decl., ¶32.) The purported novelty of the ’388 patent
`
`therefore lies in the use of correction information, generated by a decoding module,
`
`- 8 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`in the step of ISI2 correction: “[t]he technique of the invention therefore consists in
`
`computing an item of correction information that represents the received symbol[,]
`
`and in taking account of one or more of these [correction] items of information to
`
`correct the [ISI].” (’388 patent, 3:58–61.)
`
`The ’388 patent describes two structures for performing iterative reception.
`
`(Xiong Decl., ¶36.) The first structure illustrated in Figure 4 (annotated below) uses
`
`a single module. In this structure, the output of the decoder is processed as
`
`correction information 𝑍𝑛 and fed directly back to ISI correction means 41
`
`(equalizer). As discussed below in the claim construction section, the “weighted
`
`output” of ISI correction means 41 and convolutive decoder 43 are simply non-
`
`binary outputs, as was well-known at the time. (Id., ¶48.) The decoder output is
`
`either output from the module as a decision 𝐷𝑘, or input into adder 44 to obtain
`correction information 𝑍𝑘. The correction information is then interleaved by
`
`interleaver 45 so that it is in the same order as the transmitted information, and fed
`
`back to ISI correction means 41 for another round of processing. (Id., ¶34.)
`
`
`2 The ’388 patent uses the term “IES,” which stands for the French
`
`“interférence entre symboles,” to denote inter-symbol interference. For ease of
`
`discussion, Marvell refers to inter-symbol interference using the conventional
`
`representation, “ISI.” (See, e.g., ’388 patent, 7:54–56.)
`
`- 9 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`The ’388 patent stresses that its receiver utilizes known techniques for both
`
`decoding and equalization. For example, the ’388 patent states that the ISI
`
`correction means can use a prior-art algorithm disclosed by Berrou and published in
`
`a 1992 French application. (’388 patent, 8:53–57, 9:34–37; Exhibit 1011.) The ’388
`
`patent explains that its decoding means operates “according to a standard technique
`
`of convolutive decoding,” and lists the same 1992 prior-art Berrou French
`
`application as describing an example decoding algorithm. (Id.)
`
`
`
`(Xiong Decl., Figure C.)
`
`The second structure depicted in Figure 6 (annotated below) utilizes a series
`
`of cascaded modules for iteration. (’388 patent, 10:51–54 (“in a modular fashion, by
`
`associating each module with each iteration”) (emphasis added).) Each module
`
`incorporates the structure disclosed in Figure 4 without the feedback. (Id., 10:66–
`
`11:1, 11:19–24, Figures 4, 7; Xiong Decl., ¶37.) Instead of routing correction
`
`information 𝑍𝑛 back to the original ISI correction means (equalizer) 41, it is routed
`
`- 10 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`to a second module containing a second ISI correction means (equalizer) and second
`
`decoder 43.
`
`
`
`(Xiong Decl., Figure D.)
`
`The annotated figure below inserts the structure of Figure 4 (without
`
`feedback) into a cascaded architecture such as depicted in Figure 6 where the output
`
`of one module forms the input of the subsequent module. (Id., ¶38.) For ease of
`
`discussion only two modules are shown.
`
`- 11 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`
`
`(Id., ¶38, Figure E.) Each cascaded module receives symbol input 𝑅𝑖, which is the
`distorted by ISI and noise. The symbol input 𝑅𝑖 is input to ISI correction means
`correction information input 𝑍𝑛 of the first module in the cascade is “set at a neutral
`correction information input 𝑍𝑛 of the second (and any subsequent) module is
`
`value” such that it “has no effect on the computations” (Id., 10:31–32), whereas the
`
`sequence of received symbols corresponding to the transmitted symbols but
`
`(equalizer) 41, although the symbol input to the second module is delayed
`
`corresponding to the latency of the first module. (’388 patent, 10:55–58.) The
`
`received from the correction information output of the first (or prior) module. (Id.,
`
`- 12 -
`
`

`

`10:55–61.) Conversely, the decoder output 𝐷𝑘 is ignored in the first module and is
`
`output from the last module in the cascade (e.g., the second module in Figure E).
`
`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`(Id., 10:61–62.)
`
`Under either architecture, the ’388 patent performs a first ISI correction,
`
`followed by a first decoding, and then performs a second iteration of ISI correction,
`
`followed by a second iteration of decoding. “The essential characteristic of the
`
`invention is the iterative association of a symbol detection module with weighted
`
`outputs and a decoding module, also with weighted outputs.” (Id., 8:44–46.) The
`
`’388 patent illustrates the steps of its iterative reception process in Figure 5
`
`(annotated below).
`
`- 13 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`
`
`
`(Xiong Decl., ¶39, Figure F.)
`
`C. Level of ordinary skill in the art.
`A person of ordinary skill in the art (POSITA) at the relevant time of the ’388
`
`patent would have at least a Bachelor’s degree in electrical engineering, computer
`
`- 14 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`science, computer engineering, or related field of study, or equivalent experience,
`
`and at least two years of experience in studying or analyzing digital equalization and
`
`coding systems. (Id., ¶22.) A POSITA would have been familiar with various types
`
`of coding techniques including turbo coding, convolutional coding, and block
`
`coding, as well as various types of equalization techniques including decision
`
`feedback equalization, linear equalization, and maximum likelihood equalization,
`
`and would have been familiar with the similarities between equalization and
`
`decoding. (Id., ¶22.)
`
`D. Claim construction.
`The ’388 patent expired February 6, 2016. 35 U.S.C. §154(a)(2). The Board
`
`should therefore apply the district court claim construction standard as articulated in
`
`Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005). Except for the constructions
`
`set forth below the terms are to be given their plain and ordinary meaning as
`
`understood by a POSITA and consistent with the disclosure.3
`
`
`3 Marvell’s proposed constructions do not constitute an admission that the
`
`claims are valid under 35 U.S.C. §112. Marvell reserves the right to challenge the
`
`patentability of any claim under §112 in other forums.
`
`- 15 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`1. The Board should construe “weighted [values/symbols]” as
`“non-binary [values/symbols].”
`Claim 1 recites “inter-symbol interference correction means…delivering
`
`estimated symbols Ai,1 with weighted values,” and “means for the decoding of said
`
`estimated symbols…delivering decoded symbols Ai,2 with weighted values.” Claim 4
`
`recites “said means to compute the item of correction information comprise
`
`means…delivering weighted symbols.” The ’388 patent does not define the terms
`
`“weighted values” or “weighted symbols,” nor are these phrases terms of art with a
`
`plain meaning. (Xiong Decl., ¶47.)
`
`The terms “weighted values”/“weighted symbols” are associated with
`
`decoding and ISI correction. The ’388 patent refers to French application FR91-
`
`05279 as disclosing an example algorithm to be used for both ISI correction and
`
`decoding. (’388 patent, 8:53–57, 10:36–39.) U.S. Patent No. 5,406,570, provided as
`
`Exhibit 1025, claims priority to that same French application. (Ex. 1025, [30].) The
`
`’570 patent discusses the “standard way” of decoding, which “gives a binary
`
`estimation of each symbol.” (Id., 1:19–22.) The ’570 patent further states that, in
`
`contrast to giving a binary estimation, “it is especially useful to weight the decisions
`
`taken by a maximum likelihood decision algorithm” (Id., 1:26–28) and describes
`
`“[a] weighted decision decoder [that] delivers an information element coded on n
`
`bits at output.” (Id., 1:34–35.) The first bit is the binary decision “identical to that
`
`delivered by the standard decoder, and the remaining n-1 bits represent[] the
`
`- 16 -
`
`

`

`Petition for Inter Partes Review of
`U.S. Patent No. 6,108,388
`reliability assigned to the decision.” (Id., 1:35–38.) This description matches the
`
`definition of the more common term “soft output” or “soft decision.” (Lin, Ex. 10