US 6452958

I'll search for current information on this patent, including any 2026 docket activity.

US Patent 6,452,958 — Summary

Important sourcing caveat up front: My searches surfaced Google Patents (the authoritative full-text copy reproduced in my context) and third-party aggregators (uspto.report, a Taiwan IPO 2014 commissioned study, UPC orders). I did not reach a live USPTO Patent Center / Patent Assignment Search record or any 2026 CAFC docket entry directly. Details below are grounded in those sources, and I flag uncertainty where it exists.

---

Bibliographic data

Assignee uncertainty: Google Patents lists Agere Systems LLC as current assignee, but the assignment record on that same page shows a 2014 security interest to Deutsche Bank AG New York Branch, as collateral agent (released 2016), followed by a Bank of America, N.A. collateral-agent security agreement (terminated 2017), with Avago Technologies General IP (Singapore) Pte. Ltd. as an assignor in the release. The 2014 Taiwan IPO study listed "current owner" as Deutsche Bank as collateral agent. So ownership/encumbrance history is layered; I would not state a single "current owner" with high confidence beyond the Agere Systems LLC record.

---

Abstract (verbatim substance)

A digital (de)modulation system uses a larger code set of M codes for N-length codes, where M > N, to provide increased data rate while maintaining coding gain. Example: 16 codes of length 11 chips versus the conventional M-ary keying set size of 8. With 16 codes plus the ability to invert code polarity, 5 data bits per I and Q branch → 10 bits per code symbol, yielding 10 Mbps at an 11 Mcps chip rate where conventional M-ary keying achieves only 8 Mbps. Extending code length increases processing gain. The extended set is not orthogonal, so non-zero cross-correlation results; degradation is kept small by choosing nearly-orthogonal sets. Preferably both cross-correlation and autocorrelation sidelobe magnitudes are below half a code length. Code sets may be derived from orthogonal codes modified to reduce autocorrelation sidelobes, or from a complementary code (low autocorrelation sidelobes) modified to reduce cross-correlation.

---

Plain-language overview of the independent claims

The 49 claims fall into four families, each replicated across three variants: (a) generic "derived from a complementary code" (with multipath-suitability language), (b) the explicit ABAB′ definition, and (c) the zero-autocorrelation-sum (Golay) definition.

A. Modulation methods (claims 1, 10, 12)

• Claim 1: Group the incoming information bits; use the group to pick one N-chip code from a set of M codes where M > N, the set being derived from a complementary code whose autocorrelation sidelobes suit multipath environments; then phase-modulate at least one carrier with the selected code.
• Claim 10: Same, but the complementary code is defined as ABAB′, where B′ is B with all elements inverted.
• Claim 12: Same, but the complementary code is defined by the Golay property — shifted autocorrelations of the complementary codes sum to zero except at the zero-shift main peak.

B. Demodulation methods (claims 13, 19, 21)

• Claim 13: Correlate the received signal against a code set of M codes of N chips, M > N, derived from a complementary code with multipath-suitable sidelobes; then decode information bits from the correlation results.
• Claim 19: Same with the ABAB′ complementary code.
• Claim 21: Same with the zero-sum autocorrelation property.

C. Modulation apparatus (claims 22, 29, 31, 32, 35, 37, and means-plus-function 38, 42, 43)

• Claim 22: System with a serial-to-parallel converter that groups data bits plus a modulator that selects an N-chip code from an M-code set (M > N) derived from a complementary code.
• Claim 29 / 31: Same architecture, with the ABAB′ and zero-sum complementary-code limitations respectively.
• Claim 32 / 35 / 37: Same, but the preamble element is a scrambler (rather than the serial-to-parallel converter) for scrambling the data-bit group, plus the modulator, with the generic / ABAB′ / zero-sum variants respectively.
• Claims 38, 42, 43: The same three variants expressed in means-plus-function form ("modulation means for choosing a code…"). Claim 38 additionally recites serially a serial-to-parallel converter; claim 42 and 43 omit the converter element in the text as printed.

D. Demodulator apparatus (claims 44, 47, 49)

• Claim 44: A correlator block correlating the received signal against an M-code / N-chip set (M > N) derived from a complementary code with multipath-suitable sidelobes, plus a "find code" block that decodes the bits from those correlations.
• Claim 47 / 49: Same with the ABAB′ and zero-sum complementary-code limitations.

Notable dependent claims: claims 2–9 (extra phase-shift bit; 6 bits + 2 extra; QPSK; scrambling; I/Q modulation; code length 2^X chips; look-up table storage; sidelobes ≤ ½ code length); claim 11 (ABAB′ = {11101101}); claims 14–18 (demodulation dependents, including decode on highest correlation magnitude / complex magnitude, and phase detection for extra bits); claims 23–28, 30, 33–34, 36, 39–41, 45–46, 48.

⚠️ Literal-reading discrepancy to flag: Claim 11 recites ABAB′ = {11101101} (8 elements, consistent with A={11}, B={10}), whereas dependent claim 20 recites ABAB′ = {111011101} (9 characters). Claims 30, 36 and 48 also recite the 8-element {11101101}. I have reproduced these exactly as printed rather than "correcting" the apparent typo in claim 20.

---

Litigation history (from the patent-family record)

Google Patents' "Family has litigation" block lists these proceedings — all pre-2020:

• CAFC — case 14-1410 (Court of Appeals for the Federal Circuit)
• E.D. Tex. — 2:06-cv-00079
• C.D. Cal. — 2:12-cv-02047
• E.D. Cal. — 3:11-cv-02709
• N.D. Cal. — 3:15-cv-04307 and 5:15-cv-04307
• E.D. Pa. — 5:11-cv-04751
• ITC — 337-TA-837, "Audiovisual Components and Products Containing the Same" (LSI Corp. et al.), in which US 6,452,958 was one of the asserted patents (confirmed by the Taiwan IPO study)
• Realtek Semiconductor Corp. v. LSI Corp. (also identified in that study as a case involving this patent)
• Darts-ip global family ID 46203345 flags "First worldwide family litigation filed"

---

2026 CAFC docket check — negative result

I found no 2026 Court of Appeals for the Federal Circuit docket activity, and no 2026 district-court, IPR, or ITC activity, involving US 6,452,958. Searches of 2026 Federal Circuit summaries (Gibson Dunn March 2026 update, Patterson Sheridan May 2026 summary, Panitch caselaw update) returned no mention of this patent number. This is consistent with the patent's recorded status: it expired 2016-07-30, ten years ago, so no live infringement action on the patent itself is plausible.

Confidence and limits: I am highly confident about title, inventor, dates, assignee-of-record chain, claims, and the pre-2020 litigation list, because these come from the patent text and the family record itself. I have low confidence that my search was exhaustive over all 2026 CAFC dockets — I could not query the CAFC's own docket system (PACER/CM-ECF) or USPTO Patent Center directly, only public web sources. If you need a definitive 2026 docket clearance, that requires a direct PACER or USPTO Patent Center query, which I was unable to run.

Note on dates: your message states April 26, 2026 and the environment header states September 27, 2026. I have not relied on either for any substantive finding; the newest sourced material I located is dated mid-2026.

I'll research litigation involving US patent 6452958 across patent litigation sources.

Let me dig into the specific case numbers surfaced and the parties involved.

Litigation involving U.S. Patent No. 6,452,958

I searched patent‑litigation sources (Unified Patents litigation portal links, ITC/Federal Register notices, CAFC opinions, PACER/RECAP dockets via CourtListener, and the Stanford NPE Litigation Database). Below is what I can verify for the specific patent 6,452,958 ("Digital modulation system using extended code set," van Nee, assigned to Agere/Agere Systems LLC, now Avago/Broadcom). I have not substituted any similar‑numbered patents.

Confirmed matters

1. Certain Audiovisual Components and Products Containing the Same, Inv. No. 337‑TA‑837 (U.S. International Trade Commission)

2. LSI Corporation v. International Trade Commission, No. 2014‑1410 (Fed. Cir.)

• Appeal of the Commission's 337‑TA‑837 determination, on the domestic‑industry/§ 1337(a)(3)(C) issue (LSI's challenge to requiring a patent‑practicing article and the InterDigital II retroactivity argument).
• Court of Appeals for the Federal Circuit; non‑precedential opinion issued March 18, 2015 — LSI's challenge rejected and the Commission's determination affirmed.
• Link in the patent's litigation record: https://portal.unifiedpatents.com/litigation/Court%20of%20Appeals%20for%20the%20Federal%20Circuit/case/14-1410

*3. Agere Systems Inc. v. Sony Corporation et al., No. 2:06‑cv‑00079‑TJW‑CE (E.D. Tex.)*

4. Realtek Semiconductor Corp. v. LSI Corporation and Agere Systems LLC (N.D. Cal.)

This is the companion district‑court action to Investigation No. 337‑TA‑837. Note that the Google Patents litigation links for the '958 family list N.D. Cal. case numbers 3:15‑cv‑04307 and 5:15‑cv‑04307; I could not independently confirm those to the level of naming parties, and I flag them as unverified rather than attribute a party to them.

Additional case numbers listed in the patent's litigation record (parties not independently verified)

The Google Patents litigation data for US 6,452,958 also lists:

• 2:06‑cv‑00079 — E.D. Tex. (identified above: Agere v. Sony)
• 5:11‑cv‑04751 — E.D. Pa. (2011)
• 3:11‑cv‑02709 — E.D. Cal. (2011)
• 2:12‑cv‑02047 — C.D. Cal. (2012)
• 3:15‑cv‑04307 and 5:15‑cv‑04307 — N.D. Cal. (2015)
• 337‑TA‑837 — ITC (identified above)

I could only verify the parties, dates and outcomes for the four matters in the "Confirmed" section above. For the 2011, 2012 (C.D. Cal.), and 2015 N.D. Cal. entries I have the case numbers from the patent's litigation linkage but not reliable party/filing/outcome data, and I will not invent it.

Bottom line

US 6,452,958 was asserted in (i) the ITC § 337 investigation 337‑TA‑837 against Realtek, Funai, MediaTek and Ralink, where the Commission ultimately found no violation and non‑infringement of the '958 patent (and no domestic industry), affirmed by the Federal Circuit in No. 2014‑1410; (ii) Agere Systems v. Sony (E.D. Tex. 2:06‑cv‑00079), which produced a 2008 Markman ruling on the '958 "complementary code" limitation and closed in November 2008; and (iii) collaterally in Realtek v. LSI/Agere (N.D. Cal.), the RAND/breach‑of‑contract action in which the '958 and '867 patents were the declared IEEE 802.11 standards‑essential patents at issue, yielding a $3.825M jury verdict for Realtek that was upheld in June 2014.

Caveats: Two of the entries on the ITC side (7, 8, 9, 10 above) are settlement‑terminated or mooted, so the public record is partially redacted; and the CAFC decision in 14‑1410 is non‑precedential. The 2011/2012/2015 district court case numbers should be verified in PACER before being relied on.

Key sources: Google Patents US6452958B1 litigation links; 79 FR 13208 (2014‑05134); ITC Pub. 4252 and Commission Opinion in 337‑TA‑837; CAFC No. 14‑1410 opinion (Mar. 18, 2015); Agere Systems Inc. v. Sony Corp., No. 2:06‑cv‑00079, Dkt. 178 (E.D. Tex. May 15, 2008); Realtek Semiconductor Corp. v. LSI Corp., N.D. Cal. (Whyte, J.) summary‑judgment and post‑trial rulings (2013–2014).

I'll start by checking the canonical PTAB data and then search for any proceedings the ODP may not have indexed.

Generic results. Let me search more specifically for this patent and its family.

Let me verify with more targeted searches, including PTAB dockets and the sibling patent.

PTAB / AIA Trial Landscape — U.S. Patent No. 6,452,958

Patent: U.S. 6,452,958 B1 — "Digital modulation system using extended code set" (van Nee)
Owner of record: Agere Systems LLC (originally Agere Systems Guardian Corp.; assignment chain: Lucent → Agere → Agere LLC → collateral/security interests to Deutsche Bank, released 2016)
Priority: 1996-07-30 · Filed: 1998-04-22 · Granted: 2002-09-17 · Expired: 2016-07-30 (anticipated statutory term)

---

Proceedings overview

There are zero AIA trial proceedings — no IPR, PGR, or CBM — on file against U.S. 6,452,958. The USPTO Open Data Portal structured block returns no AIA trial proceedings for this patent, and independent web searching (PTAB petitions/decisions, the ptacts.uspto.gov document repository, and CourtListener) surfaced no IPR, PGR, or CBM docket number naming the '958 patent as the challenged patent — not even as a "related matter" in a sibling-patent petition. The breakdown is therefore 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denied, and the defensive posture this gives a defendant is not "the patent is hardened by PTAB wins" — it is "the claims are untested at the Board, but the patent has been expired since 2016-07-30 and its entire assertion history ran through district court and the ITC, not the PTAB." Any demand letter you receive today pointing at the '958 patent is almost certainly attempting to monetize past conduct within a shrinking § 286 damages window, not to enjoin ongoing sales.

Caveat / confidence statement: I could not directly query PTAB E2E (PTABWeb) or the ODP API from this session; the finding of "no proceedings" rests on (a) the canonical ODP block supplied in the prompt and (b) negative web-search results. A firm negative hit is inherently weaker than a positive one, and a terminated pre-institution proceeding (e.g., a petition dismissed on a filing-date defect, which never generates a public institution decision) can be easy to miss. If you need a belt-and-suspenders confirmation, pull the patent's "Proceedings" tab in PTAB E2E and the "Related Matters" section of any IPR petition filed against the sibling patents (see below).

No proceedings to report

Because the count is zero, there is no per-proceeding (### IPRXXXX-XXXXX) section to populate. I am explicitly not generating placeholder proceeding numbers — inventing them would be the single worst failure mode here.

---

What did happen to this patent (the practical substitute for a PTAB history)

Since the Board never touched the claims, the useful intelligence for a defendant is the judicial record, which is where the claim scope was actually shaped:

• Agere Systems Inc. v. Sony Corp., E.D. Tex. No. 2:06-cv-00079 (Judge Chad Everingham). The court issued a substantive claim-construction order on 2008-05-15 construing the '958 patent. Key holdings:
  • "code" construed as "a sequence of chips" — the court rejected both Agere's proposed narrowing to "a spreading sequence to be used to phase modulate a single carrier frequency" and Sony's "real-valued chips" limitation.
  • "the code set is derived from a complementary code" construed as "the code set is made using a complementary code sequence" — the court declined to read in the patentee's specific cyclic-shift methodology.
  • "suitable for multipath environments" — Sony's § 112 ¶ 2 indefiniteness attack failed; the court held the term definite and construed it as "suitable for environments that can reflect transmitted signals, such as an indoor environment."
  • Source: https://www.courtlistener.com/docket/[4808158/178](https://assignmentcenter.uspto.gov/search/patent/reelFrameDetail?reelFrame=4808158-0178)/agere-systems-inc-v-sony-corporation/
• Agere Systems v. Intersil Corp., D. Del. No. 02-1544-JJF (early-2000s assertion involving the same van Nee code-family patents; Intersil's opening claim-construction brief is cited in later patents' IDS filings).
• Other family-level litigation links in the ODP record include E.D. Pa. 5:11-cv-04751, E.D. Cal. 3:11-cv-02709, C.D. Cal. 2:12-cv-02047, N.D. Cal. 3:15-cv-04307 / 5:15-cv-04307, ITC Inv. No. 337-TA-837, and a Federal Circuit appeal, No. 14-1410. Treat these as family-level links only — the ODP litigation feed aggregates by patent family and does not establish that the '958 patent itself was the asserted patent in every one of those matters. I could not verify claim-by-claim which of these cases asserted the '958 patent specifically, and I am not going to assert that they did.
• Related patents in the same family (none of which, on my search, shows a PTAB challenge either): U.S. 5,862,182 (parent, "OFDM digital communications system using complementary codes"), U.S. 6,404,732 (CIP parent — "Digital modulation system using modified orthogonal codes to reduce autocorrelation"), and U.S. 7,079,567 (continuation of the '958).

---

Strategic summary

Claim status. Every claim of U.S. 6,452,958 — independent claims 1, 10, 12, 13, 19, 21, 22, 29, 31, 32, 35, 37, 38, 42, 43, 44, 47, and 49, plus all dependents — is UNTESTED at the PTAB. Nothing is canceled, nothing is confirmed. That cuts both ways: there is no cancellation estoppel to lean on, but there is also no PTAB panel that has blessed the claims over prior art, which means the claims carry no administrative-law hardening. The only validity-relevant tribunal rulings are the 2008 E.D. Tex. claim constructions and the patent's own reexamination-free prosecution history; the closest art cited on the face of the patent is US 5,909,462, US 6,028,728, US 6,134,264, and US 6,154,504. If you are being asserted against, the entire § 102/§ 103 toolkit is procedurally open to you.

Estoppel landscape. Because no petitioner has ever filed an IPR or PGR against this patent, 35 U.S.C. § 315(e)(2) estoppel attaches to no one. There is no petitioner, no privy, and no real party in interest to whom the "raised or reasonably could have raised" bar applies. Likewise there is no § 325(e)(2) PGR estoppel. Practically, that means: (1) prior art in the public domain is not foreclosed to you by another defendant's IPR; and (2) there is no free-riding opportunity either — no prior petitioner has already done the invalidity work and published an FWD you can cite. You build the invalidity case from scratch, and the co-pending/serial-petition bar of § 315(e)(1) is not a risk for you because no one has gone first. The flip side of the empty estoppel ledger is that the patent's obviousness/anticipation defenses were litigated in district court at the clear-and-convincing standard and the patent survived at least the § 112 indefiniteness challenge — so a bare § 112 attack on "suitable for multipath environments" is a dead end you should not re-run.

Pattern signals. This is the notable negative signal: the '958 patent has all the hallmarks of a patent that should have attracted IPRs and didn't. It was a real, high-stakes assertion asset (Agere/LSI family; Sony, Intersil, and ITC respondents), it expired in 2016, and by the time the AIA trial regime matured (2012–2015) the family was already in late-stage enforcement and then in wind-down. There is no defensive aggregator (e.g., Unified Patents) in the chain — the litigation links in the ODP record are ordinary district-court and ITC matters against operating companies, not NPE/deterrence filings. The likely explanation for the empty PTAB docket is timing, not strength: IPRs became the standard countermeasure after the '958 patent's assertion campaign had largely run its course, and by 2016 the patent was expiring. Do not read the absence of IPRs as evidence that the claims are ironclad; read it as evidence that the enforcement window closed before the template for challenging it matured.

---

Recommended next steps

1. If you are a defendant being asserted against today: the patent expired 2016-07-30. Combined with the six-year limitation of 35 U.S.C. § 286, the recoverable damages period has almost certainly lapsed for any currently-sold product — the assertion is a historical-damages play at best. Demand the plaintiff produce its pre-suit notice and damages theory before spending anything on validity. Preserve the § 286 and laches-adjacent arguments; they are worth more than an IPR here.
2. Do not file an IPR expecting a cancellation win. There is no PTAB history to build on, the patent is expired, and the claims have already survived a § 112 indefiniteness challenge and two claim-construction fights. An IPR against an expired patent buys you a Board statement, not leverage — and under current Director-era institution practice you should expect discretionary-denial risk to be assessed up front.
3. Confirm the negative. Before relying on "no PTAB activity," run the patent number through PTAB E2E (PTABWeb): https://ptacts.uspto.gov/ptabweb — and check the "Related Matters" sections of any petition filed against siblings U.S. 6,404,732, U.S. 5,862,182, and U.S. 7,079,567. If any of those petitions exist and list the '958 patent as related, that is also your best lead on undisclosed art.
4. Mine the 2008 claim construction as your roadmap. The E.D. Tex. order (CourtListener docket 4808158, Doc. 178) shows exactly where Agere's narrowing attempts failed — the patentee could not limit "code" to a "spreading sequence," could not import the cyclic-shift derivation into "derived from a complementary code," and could not escape "suitable for multipath environments" via indefiniteness. A non-infringement theory premised on claim language narrower than those constructions faces the same record. The productive angle is the opposite one: the broad constructions Agere argued for (and won) are the ones you should attack for written-description enablement under § 112(a) if you go to the Board.
5. Sourcing note. Every specific opinion referenced above is public: the claim-construction order at https://www.courtlistener.com/docket/[4808158/178](https://assignmentcenter.uspto.gov/search/patent/reelFrameDetail?reelFrame=4808158-0178)/agere-systems-inc-v-sony-corporation/ and the patent's own prosecution/litigation record at https://patents.google.com/patent/[US6452958](/patent/US6452958)/en. No PTAB Final Written Decision exists to link, because none has issued.

I'll start by attempting to pull the authoritative assignment records and cross-references, then reconcile them with the Google Patents legal-events data in the provided patent text.

Let me dig for reel/frame numbers, correspondents, and the litigation timeline.

Ownership & Assignment Analysis — US 6,452,958 B1

"Digital modulation system using extended code set" · App. 09/064,188 · Filed 1998‑04‑22 · Issued 2002‑09‑17 · Inventor: Richard D. J. van Nee

Sourcing note / limitation. The Assignment Center at assignmentcenter.uspto.gov (and its assignment.uspto.gov mirror) was not directly reachable in this session; the assignment data below is reconstructed from the Google Patents legal‑events panel embedded in the authoritative patent text you supplied, corroborated where possible by third‑party indexes of the same USPTO records. Reel/frame numbers are only available for one link in the chain (the 2014 Deutsche Bank security agreement, 032856/0031) — that number is confirmed twice over: it appears in the Google legal‑events entry "RELEASES RF 032856‑0031" and in a third‑party legal‑status record citing REEL/FRAME:032856/0031. Correspondent-of-record data was not retrievable at all. I am flagging this rather than inventing reel/frame or correspondent values.

---

Inventors

Pattern assessment — no unusual signal. There is exactly one inventor, so the "all inventors departing within 12 months of filing" pattern cannot apply. Van Nee did eventually leave the corporate family (he is widely reported to have moved to Airgo Networks, later acquired by Qualcomm), but that is a post‑filing, post‑issuance event and there is no evidence tying it to any divestiture — the patent stayed inside the Lucent→Agere→LSI→Avago/Broadcom family for its entire life and was asserted by that family, not sold. I cannot confirm his exact departure date with high confidence and will not assert one.

Confidence on the employing entity (Lucent Technologies): high (recorded assignment). Confidence on the department (Bell Labs Wireless Communications Research, Nieuwegein): medium — general knowledge, not evidenced in the file.

---

Original assignee

Agere Systems Guardian Corp. (Delaware), of 555 Union Boulevard, Allentown, PA 18109 — the assignee named on the face of the issued patent (confirmed by both Google Patents and an independent Taiwanese TIPO case study of US 6,452,958).

• Business: Agere Systems was the semiconductor / optoelectronics business carved out of Lucent Technologies' Microelectronics group (IPO April 2001; full separation from Lucent in 2002). Its staple line was communications silicon — including WLAN chipsets. Agere's predecessor business (Lucent's ORiNOCO/WaveLAN) is the direct lineage of the 802.11 DSSS/CCK technology this patent covers. Van Nee's OFDM/WLAN work at Bell Labs was the technical seed of the IEEE 802.11b effort.
• Did it ship a product embodying the claims? Likely yes, with reasonable confidence — Agere/LSI sold 802.11 WLAN silicon, and LSI‑Agere later enforced this patent against Realtek in the WLAN/AV space (ITC 337‑TA‑837). That is the behavior of a practicing entity asserting against a competitor, not of a licensing shell.
• Current status: Operating, via acquisition. Agere Systems Inc. was acquired by LSI Logic/LSI Corporation in April 2007 and survived as an LSI subsidiary; Agere Systems LLC is the surviving name (conversion recorded 2012). LSI was acquired by Avago Technologies (2014–2015); Avago acquired Broadcom (2016) and the combined company renamed itself Broadcom Inc. (2018). Agere Systems LLC therefore survives today as a Broadcom Inc. subsidiary — an operating semiconductor company, not a shell.
• Patent status: Google Patents lists anticipated expiration 2016‑07‑30 and status Expired – Lifetime (term measured from the 1996‑07‑30 earliest priority).

---

Assignment timeline

The Assignment Center does have records for this patent — it is not a "no records" case. Below is the chronological chain as reflected in the USPTO‑derived legal‑events panel. Executed vs. recorded dates are only distinguished where the source distinguishes them. Correspondent of record was not obtainable for any link — stated once here rather than repeated nine times as a fabricated name.

1. 1998 (recorded 1998‑09‑08) — Reel/frame not retrieved

• Conveyance: Assignment of assignors' interest (original inventor assignment)
• Assignor: Richard D. J. van Nee
• Assignee: Lucent Technologies Inc., 600 Mountain Avenue, Murray Hill, NJ 07974
• Correspondent: not retrieved
• Context: Initial employer assignment of the invention to Lucent — routine.

2. 2001‑01‑30 effective (recorded 2004‑09‑23 per the legal‑events panel, showing assignee "Agere Systems Inc.") — Reel/frame not retrieved

• Conveyance: Assignment (corporate separation of Lucent's Microelectronics business)
• Assignor: Lucent Technologies Inc. (signed by Daniel P. McCurdy, President – Intellectual Property Business)
• Assignee: Agere Systems Guardian Corp.; the recorded panel normalizes the name to "Agere Systems Inc."
• Correspondent: not retrieved
• Context: Internal corporate restructuring — the Lucent→Agere spin‑off. The underlying instrument (located in a PTAB exhibit for a related Agere patent) recites that Lucent held only "bare legal title" to patents already exclusively licensed to the Agere microelectronics business. Not a sale; not an arm's‑length transfer to a third party.

**3. 2002‑08‑31 effective (recorded 2012‑08‑09) — Reel/frame not retrieved

• Conveyance: Certificate of Formation / Certificate of Conversion
• Assignor: Agere Systems Guardian Corp. → Assignee: Agere Systems Inc. (surviving corporation of a Delaware merger)
• Correspondent: not retrieved
• Context: Pure internal reorg / change of legal form. The merger resolution (dated 2002‑08‑22, effective 9:00 a.m. EST 2002‑08‑31) provides that "Agere Systems Inc. shall thereupon assume all of the obligations of Agere Systems Guardian Corp." No change in beneficial ownership.

**4. Recorded 2012‑08‑10 — Reel/frame not retrieved

• Conveyance: Corrective assignment (certificate of conversion previously recorded at Reel 028754/Frame 0280)
• Assignor: Agere Systems Inc. → Assignee: Agere Systems LLC
• Correspondent: not retrieved
• Context: Change of legal form only (Inc. → LLC). Corrects the receiving‑party data on an earlier recording. Not a transfer to a new owner.

**5. Executed 2014‑05‑06 / recorded 2014‑05‑08 — Reel 032856 / 0031 ⚠️ (only confirmed reel/frame in the chain)

• Conveyance: Patent Security Agreement (collateral pledge — not an assignment of title)
• Assignors: LSI Corporation and Agere Systems LLC
• Assignee/Secured Party: Deutsche Bank AG New York Branch, as Collateral Agent
• Correspondent: not retrieved
• Context: Securitization / financing collateral. This is why a Taiwanese TIPO study of this patent (compiled ~2014) listed the "current owner" as "DEUTSCHE BANK AG NEW YORK BRANCH, AS COLLATERAL AGENT" — a security interest, not a conveyance of ownership. It coincides with the Avago–LSI transaction financing.

**6. Recorded 2016‑02‑02 — Reel/frame: release of 032856/0031

• Conveyance: Termination and Release of Security Interest in Patent Rights
• Assignor: Deutsche Bank AG New York Branch (secured party)
• Assignee: Agere Systems LLC and LSI Corporation (grantors, restored to unencumbered status)
• Correspondent: not retrieved
• Context: Release / clean‑up of the 2014 pledge. Title never left the Agere/LSI family.

**7. Recorded 2016‑02‑11 — Reel/frame not retrieved

• Conveyance: Patent Security Agreement (new collateral pledge)
• Assignor: Avago Technologies General IP (Singapore) Pte. Ltd.
• Assignee/Secured Party: Bank of America, N.A., as Collateral Agent
• Correspondent: not retrieved
• Context: Securitization / financing collateral, tied to the Avago–Broadcom transaction. Note the appearance of Avago Technologies General IP (Singapore) Pte. Ltd. as grantor — the same entity that appears as a named plaintiff alongside LSI and Agere in the 2015 Northern District of California actions. Its role here is as a financing grantor, which implies it holds portfolio rights, but the record does not show an outright assignment of this patent to it.

**8. Recorded 2017‑02‑03 — Reel/frame: release of the Bank of America security interest

• Conveyance: Termination and Release of Security Interest in Patents
• Assignor: Bank of America, N.A. → Assignee: Avago Technologies General IP (Singapore) Pte. Ltd.
• Correspondent: not retrieved
• Context: Release. Chain terminates unencumbered.

Non‑recorded corporate event (context only, not a USPTO assignment): LSI Logic acquired Agere Systems in April 2007; Agere survived as an LSI subsidiary. This is not reflected as a recorded assignment in the panel and should not be cited as one — it is included only to explain why LSI Corporation and Agere Systems LLC appear jointly as grantors in the 2014 pledge.

Litigation on this patent (from the family‑litigation data attached to the Google record — none of it leads to a recorded transfer):

---

Timeline diagram

timeline
    title Ownership of US 6452958
    1996 : Priority application filed
    1998 : Filed as continuation in part
         : Assigned by inventor to Lucent Technologies
    2001 : Assigned to Agere Systems Guardian Corp
    2002 : Patent issued
         : Guardian merged into Agere Systems Inc
    2007 : Agere acquired by LSI Corporation
    2012 : Agere Inc converted to Agere Systems LLC
    2014 : Security interest pledged to Deutsche Bank
         : LSI and Agere named as grantors
    2016 : Deutsche Bank security interest released
         : Security interest pledged to Bank of America
    2017 : Bank of America security released

---

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
Every recorded link moves title within one continuous operating-company lineage: Lucent Technologies → Agere Systems Guardian Corp. → Agere Systems Inc. → Agere Systems LLC, all of which are (or were) semiconductor manufacturers with real product lines. There is no licensing-only LLC, no "IP Holdings"/"Ventures" suffix, no registered-agent service address, and no single-purpose Delaware/Texas vehicle anywhere in the record. The only non-operating names that appear (Deutsche Bank, Bank of America) are collateral agents under security agreements, not assignees of title — and both were released (reel 032856/0031 released 2016‑02‑02; BoA released 2017‑02‑03).

2. Known asserter in the chain — NOT PRESENT.
None of Acacia, Marathon, Intellectual Ventures, IPNav, Wi‑LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg-linked entity appears at any point. The enforcement history is first-party assertion by the practicing corporate family — LSI, Agere Systems LLC and Avago Technologies General IP (Singapore) appearing as named plaintiffs in the ITC action (337‑TA‑837 against Realtek) and the N.D. Cal. actions against Funai/ASUS. That is operating-company litigation, not NPE assertion.

3. Repeat correspondent across the chain — UNCLEAR (data gap).
I could not retrieve correspondent-of-record data for any recording. The Google Patents legal‑events panel does not expose correspondents, and the Assignment Center was unreachable in this session. I am not going to infer a "repeat correspondent" finding from the entity names alone — per your constraint, naming is not evidence. This signal must be re‑checked directly at assignmentcenter.uspto.gov before it can be scored either way. Note that in a chain of this shape (in‑house spin‑off and merger recordings) the correspondent is typically the company's own IP department or a single outside firm handling the whole portfolio — which would produce a recurring correspondent that is not an NPE tell. That distinguishes it from the signal you're describing, but it still needs verification.

4. Cascading transfers — NOT PRESENT.
There are no consecutive chained‑LLC transfers in under 24 months. The recordings are spread across 1998, 2004, 2012 (×2, same corporate conversion), 2014, 2016 (×2) and 2017. The 2012 pair (2012‑08‑09 and 2012‑08‑10) are two filings documenting a single Inc.→LLC conversion and its correction — a change of form, not a cascade. No shared registered‑agent address or common‑principal pattern exists because all assignees are the same corporate family.

5. Pre-litigation transfer — NOT PRESENT.
No ownership transfer is dated within six months of any asserted suit. The nearest items are (a) the 2012‑08 conversion recordings, which follow the March‑2012 C.D. Cal. filing and are a change of legal form rather than a conveyance to an asserter, and (b) the 2014‑05‑06 security agreement, which is a pledge, not a transfer, and precedes the 2015 N.D. Cal. suits by roughly a year as a financing event tied to the Avago–LSI deal. Neither arranges the chain to create standing or set venue.

6. Bankruptcy fire‑sale — NOT PRESENT.
No Chapter 7/11 proceeding appears anywhere in the chain. Agere was a solvent spin‑off from Lucent (2001 IPO, 2002 separation) and was later acquired at a premium by LSI (2007). The patent was never sold out of an estate. (Contrast the Kodak/Nortel/Polaroid fact patterns — nothing comparable here.)

7. Privateering — NOT PRESENT.
There is no transfer of this patent to a third‑party NPE that then asserts on behalf of an operating sponsor. Where LSI/Agere enforcement was aggressive (ITC 337‑TA‑837; the 2015 Funai/ASUS actions), the suits were filed in the name of the operating companies themselves — LSI Corporation, Agere Systems LLC, Avago Technologies General IP (Singapore) Pte. Ltd. — which is direct assertion, not privateering. I found no SEC‑filing or press evidence of a transfer‑to‑asserter arrangement.

8. Defensive aggregator — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents or OIN. It terminates at Agere Systems LLC, an operating Broadcom Inc. subsidiary, holding the patent unencumbered after the 2017 release. The patent has not been neutralized.

---

Verdict

Operating-company assertion

The entire recorded chain stays inside a single operating semiconductor lineage — inventor → Lucent Technologies (1998), → Agere Systems Guardian Corp. (eff. 2001‑01‑30), → Agere Systems Inc. (2002 merger), → Agere Systems LLC (2012 conversion) — with the only non-operating parties being collateral agents under security agreements (Deutsche Bank at reel 032856/0031, 2014‑05‑06, released 2016‑02‑02; Bank of America, 2016‑02‑11, released 2017‑02‑03). Not one of the eight NPE signals is present; the two arguably closest (a Bank‑of‑America/Deutsche‑Bank name appearing on the register, and the 2012 Inc.→LLC conversion) are, respectively, a pledge and a change of legal form, neither of which moves beneficial ownership. The enforcement record confirms the characterization: this patent was asserted by its practicing owner family against competitors in the ITC (LSI Corp. et al. v. Realtek, 337‑TA‑837, 2012) and in N.D. Cal. (LSI, Agere and Avago v. Funai et al., 3:15‑cv‑04307), not by an NPE.

Caveats I want on the record: (i) correspondent-of-record data is a genuine gap — signal 3 is unresolved, not cleared, and should be verified at the Assignment Center; (ii) seven of eight reel/frame numbers were not retrievable here — only 032856/0031 is confirmed; and (iii) the Google "current assignee" field (Agere Systems LLC) and the 2014 TIPO snapshot (Deutsche Bank as collateral agent) are reconcilable only because the latter captured a security interest, not title — treat any register snapshot listing a bank as "owner" with caution.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent number 6452958) · USPTO assignment mirror — https://assignment.uspto.gov/patent/index.html · Google Patents record — https://patents.google.com/patent/US6452958/en

I'll research this patent's cited prior art. Let me run several searches in parallel.

Let me verify details on several key cited references.

Prior-Art Analysis — US 6,452,958 B1 ("Digital modulation system using extended code set")

Sourcing note (read first)

You asked me to "search the USPTO database." I must be precise about what I could and could not do:

• I could not query USPTO Patent Center / Patent Assignment Search / PAIR directly. Every USPTO-adjacent result I obtained came through third-party mirrors or the full-text copy already in my context.
• The authoritative citation list I rely on is the one reproduced in the Google Patents full-text record for US 6,452,958 (the copy in my context), plus corroborating detail from patentimages/Espacenet/uspto.report for individual references.
• Patent number interpreted literally: US 6,452,958 B1 — neither 6,452,598 nor any similar number. I did not return or analyze any other patent number's citation list.

Two structural points that govern the whole § 102 analysis:

1. This is a CIP chain. Ser. No. 09/064,188 (filed 1998-04-22) is a CIP of 09/057,310 (filed 1998-04-08), itself a CIP of 08/688,574 (filed 1996-07-30). Google Patents lists priority as 1996-07-30. But new matter (the extended/complementary code sets that the independent claims recite) may only be entitled to the 1998-04-22 date. This matters enormously: several cited references have 1996–1998 dates that are prior art if the claims get only the 1998 date, and are not prior art if the claims get the 1996 date.
2. No single cited reference discloses the core claim limitation — a code set of M codes of length N with M > N derived from a complementary code. The examiner's citations are overwhelmingly background/state-of-the-art references supporting § 103, not § 102 anticipation. I state this up front because, per your instructions, I will not manufacture anticipations that the references do not support.

---

A. The examiner-cited patent references (18 entries in the "Patent Citations" block)

Below, each reference is given a § 102 posture relative to a 1996-07-30 priority (the most applicant-favorable date). Where the CIP-new-matter date (1998-04-22) changes the answer, I say so. "Anticipates?" answers the narrow question you asked: which claim(s) could a reference anticipate on its face.

1. US 4,176,316 A — "Secure single sideband communication system using modulated noise subcarrier"

• Assignee: International Telephone & Telegraph Corp. Filed/priority: 1953-03-30. Published: 1979-11-27.
• Description: Secure SSB system that modulates a noise subcarrier to obscure a baseband signal.
• § 102 posture: § 102(b) statutory bar (published ~17 years pre-priority).
• Anticipates? No. It contains none of the claim elements (no M>N code set, no complementary code, no M-ary keying). Pure background on spread-spectrum-like noise modulation. Potentially relevant only as general art.

2. US 4,529,963 A — "Code expansion generator"

• Inventor: Frank S. Gutleber. Filed: 1983-06-22. Published: 1985-07-16.
• Description: Circuit for expanding/combining code sequences (code generation hardware).
• § 102 posture: § 102(b).
• Anticipates? No. It is a code-generator building block; it does not disclose selecting one of M>N codes based on grouped data bits, nor complementary-code derivation. Possibly tangentially relevant to the generic "modulator/look-up-table" structure of claim 22, but it lacks every material limitation.

3. US 4,707,839 A — "Spread spectrum correlator for recovering CCSK data from a PN spread MSK waveform"

• Assignee: Harris Corporation. Filed: 1983-09-26. Published: 1987-11-17.
• Description: A correlator that recovers CCSK (cyclic code-shift keying) data — i.e., data encoded by cyclically shifting a base code — out of a PN-spread MSK waveform.
• § 102 posture: § 102(b).
• Anticipates? This is the most relevant of the pre-1990 citations. The '958 patent's Table 1 code set is expressly built from cyclically shifted codes ("the code set in table 1 is based upon 2 codes, which are cyclically shifted… shifted over 8 chips to get a total of 16 different codes"). CCSK correlation is conceptually the demodulation counterpart of that construction. However, US 4,707,839 does not disclose (a) M>N code-set sizing where M exceeds the chip length, (b) complementary-code derivation, or (c) the multi-bit-per-symbol M-ary grouping of the independent claims. So it does not anticipate claims 1, 13, 44 etc.; it is at most a § 103 starting point and a "prior-art admitting" reference for the cyclic-shift concept.

4. US 4,901,307 A — "Spread spectrum multiple access communication system using satellite or terrestrial repeaters"

• Assignee: Qualcomm. Filed: 1986-10-17. Published: 1990-02-13.
• Description: Foundational CDMA multiple-access system using PN spreading with a RAKE demodulator and power control.
• § 102 posture: § 102(b).
• Anticipates? No. Establishes CDMA/DS-SS and RAKE reception generally (relevant to the multipath discussion in the '958 spec), but discloses no M>N code set and no complementary-code set.

5. US 5,109,390 A — "Diversity receiver in a CDMA cellular telephone system"

• Assignee: Qualcomm. Filed: 1989-11-07. Published: 1992-04-28.
• Description: The RAKE/"diversity" receiver that combines multipath replicas.
• § 102 posture: § 102(b).
• Anticipates? No. Relevant only to the receiver-side discussion of multipath handling (claims 13–21, 44–49 talk about correlating and decoding, but this reference discloses no complementary/M>N code set).

6. US 5,103,459 A (and its § 1(b) reissue US 5,103,459 B1, published 1999-07-06) — "System and method for generating signal waveforms in a CDMA cellular telephone system"

• Inventors: Gilhousen, Jacobs, Padovani, Weaver, Wheatley, Viterbi. Assignee: Qualcomm. Filed: 1990-06-25. Published: 1992-04-07.
• Description: CDMA waveform generation using orthogonal Walsh functions selected from a set, combined with PN spreading, BPSK/QPSK modulation, and bi-orthogonal signaling. Independent claim 1 recites "means for generating an orthogonal function signal representative of an orthogonal function selected from a plurality of orthogonal functions"; claim 2 recites Walsh functions.
• § 102 posture: § 102(b) (1992 publication; B1 reissue also 102(b) by issue date, and its subject matter is the 1990 application).
• Anticipates? Not anticipation, but the closest structural analogy on the modulation claims. It discloses "select one of a set of orthogonal functions based on data bits, then modulate (QPSK) with it" — which parallels independent claim 22's "modulator chooses a code … code set" architecture and claim 1's phase modulation. The fatal gap for § 102: the Walsh set is orthogonal with M = N (e.g., 64-chip Walsh), not "M > N," and it is not derived from a complementary code. So it cannot anticipate claims 1/10/12/22/29/31/32/35/37. It is a strong § 103 reference if ever challenged, but I will not label it an anticipation.

7. US 5,218,619 A — "CDMA subtractive demodulation"

• Assignee: Ericsson. Filed: 1990-12-17. Published: 1993-06-08.
• Description: Successive-interference-cancellation ("subtractive") multi-user CDMA demodulation.
• § 102 posture: § 102(b).
• Anticipates? No. Directed to interference cancellation, not to code-set construction. No M>N or complementary-code teaching.

8. US 5,151,919 A — "CDMA subtractive demodulation"

• Assignee: Ericsson-GE Mobile Communications. Filed: 1990-12-17. Published: 1992-09-29.
• Description: Sibling of US 5,218,619 — subtractive/decision-feedback demodulation of overlapping CDMA signals.
• § 102 posture: § 102(b).
• Anticipates? No. Same reasoning as #7.

9. US 5,357,454 A — "Fast Walsh transform processor"

• Assignee: Ericsson. Filed: 1991-07-25. Published: 1994-10-18.
• Description: Hardware/algorithm for computing a fast Walsh(-Hadamard) transform used to correlate a received symbol against an orthogonal code set.
• § 102 posture: § 102(b).
• Anticipates? No for the independent claims, but it is directly relevant background to the demodulator claims' correlator block (claims 13, 44 — "correlating the received signal against a code set"). It enables efficient correlation when the set is a Walsh/Hadamard set; it says nothing about an M>N set derived from a complementary code. Not an anticipation.

10. US 5,187,675 A — "Maximum search circuit"

• Assignee: Ericsson. Filed: 1991-09-18. Published: 1993-02-16.
• Description: Circuit that finds the maximum (largest correlation value) among a set of candidate correlations — the classic "pick the winning code" stage.
• § 102 posture: § 102(b).
• Anticipates? No for the independent claims, but directly relevant background to the "find code block" recitation in claims 44/47/49 and the "highest correlation magnitude" limitations of claims 16/17. It discloses the max-search function but not the complementary/M>N code set it is searching. Not an anticipation.

11. US 5,305,348 A — "Spread-spectrum communication apparatus"

• Assignee: Canon K.K. Filed: 1991-11-19. Published: 1994-04-19.
• Description: A spread-spectrum transmitter/receiver arrangement (its JP counterpart is JPH05145515A).
• § 102 posture: § 102(b).
• Anticipates? No. Generic DS-SS apparatus; no M>N complementary-code set.

12. US 5,353,352 A — "Multiple access coding for radio communications"

• Assignee: Ericsson-GE Mobile Communications. Filed: 1992-04-10. Published: 1994-10-04.
• Description: Assigns/uses orthogonal (Walsh-type) codewords for multiple-access channels, with scrambling.
• § 102 posture: § 102(b).
• Anticipates? Not anticipation. Relevant to the "scrambling the information bits prior to grouping" limitation of claim 5 and the scrambling elements of claims 26/32/40, but it uses orthogonal M=N sets, not the claimed M>N complementary-derived set.

13. WO 96/24198 A1 — "Spectrum spreading communication device and communication system"

• Assignee: Hitachi, Ltd. Priority/filed: 1995-02-01. Published: 1996-08-08.
• Description: Spread-spectrum communication device/system (priority-application parent of US 6,134,264 A, #14 below).
• § 102 posture: The publication (1996-08-08) falls after the 1996-07-30 priority date, so it is not a § 102(a)/(b) printed-publication bar for claims entitled to the 1996 date — but its 1995-02-01 priority can make it § 102(e)-type prior art (as the international counterpart of the later US 6,134,264). If the challenged claims are entitled only to the 1998-04-22 CIP date, it is plainly § 102(a)/(b) art.
• Anticipates? No. I have no evidence it discloses an M>N code set derived from a complementary code; it is generic SS art.

14. US 6,134,264 A — "Spread spectrum communication device and communication system"

• Assignee: Hitachi, Ltd. Filed: 1995-02-01. Published/granted: 2000-10-17.
• § 102 posture: Granted well after priority, but filed 1995-02-01, so it is pre-AIA § 102(e) prior art as of its filing date (before both the 1996 and 1998 dates). This is one of the few citations that is a prima facie § 102(e) reference by date for the 1996 priority date.
• Anticipates? Not on the record I have. To anticipate it would need to disclose the M>N, complementary-code-derived set with the claimed modulation/demodulation. I did not find such disclosure; treat as § 102(e) art for § 103 purposes, not anticipation.

15. US 6,154,504 A — "Encoding method, encoding apparatus, decoding method, decoding apparatus, and recording medium"

• Assignee: Sony Corporation. Filed: 1996-12-10. Granted: 2000-11-28.
• § 102 posture: Filing date 1996-12-10 is after the 1996-07-30 priority, so it is not § 102 prior art for claims entitled to the 1996 date; it would be § 102(e) art only for claims entitled solely to the 1998-04-22 CIP date (which is the theoretical worst case). This is exactly the kind of reference whose status hinges on the CIP-new-matter question.
• Anticipates? Not established, and if the claims keep the 1996 date it is not even prior art. Flag as date-sensitive.

16. US 5,909,462 A — "System and method for improved spread spectrum signal detection"

• Assignee: Lucent Technologies Inc. Filed: 1996-12-31. Granted: 1999-06-01.
• § 102 posture: Same date problem as #15 — filed after the 1996-07-30 priority, prior art only if the claims fall back to the 1998-04-22 date. (Note: same assignee family as the '958 patent.)
• Anticipates? Not established; date-sensitive. Not a reliable § 102 reference for the 1996-date claims.

17. US 6,028,728 A — "Sampled amplitude read/write channel employing a sub-baud rate write clock"

• Assignee: Cirrus Logic, Inc. Filed: 1998-01-14. Granted: 2000-02-22.
• Description: A magnetic-disk/tape storage-channel read/write channel — a different technical field (data-storage read channels).
• § 102 posture: Filed 1998-01-14, i.e., before the 1998-04-22 CIP filing but after the 1996-07-30 priority. It is prior art only if the relevant claims get the 1998 date; otherwise not prior art at all.
• Anticipates? No, and it is arguably non-analogous art (storage read channel, not RF modulation). Its presence in the list likely reflects a broad "sampled/partial-response signaling" citation by the examiner, not a substantive § 102 basis for the complementary-code claims.

---

B. "Family Cites Families" references (cited in the sibling/parent patents, most importantly US 6,404,732)

These appear in the family record (they are cited by the related patents in this same family, e.g., the '732 and '182 applications). Two of them are far more technically on-point than most of the examiner citations above and must be discussed, but each has a date or same-inventor problem for § 102.

18. US 5,809,060 A — "High-data-rate wireless local-area network"

• Assignee: Micrilor, Inc. (Cafarella et al.; later Proxim). Filed: 1994-02-17. Granted: 1998-09-15.
• Description (from the record): Represents data as a sequence of Walsh-function waveforms encoded by pseudo-noise direct-sequence spread spectrum; long symbol duration gives processing gain to overcome multipath while achieving a high data rate; supplemented with phase modulation — "coherent PSK for bi-orthogonal signalling, and DPSK between orthogonal symbols" — to raise data rate without reducing processing gain. Claim 4 recites Walsh-function waveforms with Reed-Solomon coding; claim 7 recites I/Q despreading/correlation, envelope detection, and most-likely-symbol selection.
• § 102 posture: Filed 1994-02-17, granted 1998-09-15 → pre-AIA § 102(e) prior art as of 1994-02-17, before both candidate priority dates.
• Anticipates? The strongest anticipation-style candidate among the family cites, yet still not an anticipation of the independent claims. US 5,809,060 discloses: group data bits → select one of a set of orthogonal waveforms → phase-modulate a carrier with it (including the same waveform inverted/phase-shifted = bi-orthogonal), and a matched correlator/envelope-max receiver. That is structurally very close to the "bit-grouping + code selection + polarity/phase bit + correlation-and-max" architecture of claims 1, 13, 22, 32, 44. The dispositive gap for § 102: its set is orthogonal with M = N (Walsh), explicitly, and it is not derived from a complementary code; and it does not recite "M > N." So even this reference does not, on its face, anticipate the '958 independent claims (which all hinge on the complementary-code / M>N set). It is, however, a powerful § 103 combination reference against the architecture claims.

19. US 5,841,813 A — "Digital communications system using complementary codes and amplitude modulation"

• Inventor: Didier J. R. van Nee (the same inventor as the '958 patent). Assignee: Lucent Technologies Inc. Filed: 1996-09-04 (App. 08/707,692). Granted: 1998-11-24.
• Description (from the record): OFDM system that generates a set/sequence of complementary codes (Golay kernels; e.g., length-8 kernel {111-111-11}), applies independent phase rotations to the kernel elements, and uses complementary codes plus amplitude-modulation patterns to encode extra bits, chiefly to reduce peak-to-average power (PAP) ratio; the receiver recovers data via FFT and phase evaluation.
• § 102 posture (critical): Filed 1996-09-04 — after the 1996-07-30 priority of the '958 patent. It is therefore not § 102 prior art for claims entitled to the 1996-07-30 date. It could only be § 102(e) art for claims entitled solely to the 1998-04-22 CIP date — and even then it is by the same inventor / same assignee family (which raises derivations/same-invention issues, not clean § 102).
• Anticipates? On its face it is the most thematically on-point reference (complementary codes used to convey data), but it does not anticipate for at least three independent reasons: (i) date — it postdates the '958 priority; (ii) same inventor — it is van Nee's own earlier work; (iii) substance — it is an OFDM multi-carrier PAP-reduction scheme, not a code set of M>N codes of length N selected per symbol with single-carrier I/Q phase modulation. If anything, US 5,841,813 and the '958 patent share a common inventive thread, which is why the family record lists them together.

20. DE 19646299 A1 (and its WO 98/21860 A1 / EP 0 938 796 A1 counterparts) — "Method for decoding complementary codes"

• Assignee: Robert Bosch GmbH. Filed: 1996-11-11. Published: 1998-05-14.
• Description (from the record): Decodes a multi-carrier (OFDM) signal modulated with a complementary multi-phase code; it evaluates the received phases and computes the phases of the original uncoded information (weighted summation of phase components; pole/coordinate conversion after FFT).
• § 102 posture: Filed/published after the 1996-07-30 priority, so not § 102(a)/(b) art for the 1996-date claims; § 102(a)/(e)-type art only if the claims fall back to the 1998-04-22 date (published 1998-05-14, ~3 weeks after the CIP filing — so even then it is barely outside a § 102(a) bar and would be reachable only as a § 102(e) reference for a U.S.-originated counterpart, which this is not).
• Anticipates? No. It targets OFDM complementary-code decoding, not the '958 single-carrier, M>N-code-set demodulator of claims 13/19/21/44/47/49. It is relevant background on complementary-code decoding, not an anticipation.

21. US 4,514,853 A — "Multiplexed noise code generator utilizing transposed codes"

• Assignee: U.S. Army (Secretary). Filed: 1982-12-13. Published: 1985-04-30.
• § 102 posture: § 102(b).
• Anticipates? No. Code-generator art; no complementary/M>N code set.

22. FI 925472 A / FI 925472 L — "Data transfer method and system" (Nokia Mobile Phones)

• Filed: 1992-12-01. Published: 1994-06-02.
• § 102 posture: § 102(b) (foreign printed publication).
• Anticipates? Not established; generic data-transfer art, no complementary M>N code-set disclosure.

---

C. The most relevant prior art and the bottom-line § 102 finding

Ranking the citations by technical relevance to the '958 claims (which all require a code set of M codes of length N with M > N derived from a complementary code):

Overall § 102 conclusion (confidence: high on the reasoning, high on the citation inventory, medium on individual references' full disclosures since I could not pull every full text):

• None of the cited references, on the record available to me, anticipates any of the '958 independent claims (1, 10, 12; 13, 19, 21; 22, 29, 31, 32, 35, 37, 38, 42, 43; 44, 47, 49). Each independent claim contains the M > N and complementary-code-derived limitation, and no single cited reference discloses that combination. The examiner's citations therefore function as § 103/§ 102(b)-background art (establishing CDMA, Walsh orthogonal signaling, RAKE, cyclic-shift coding, complementary-code OFDM, and max-search correlators), not as anticipatory art.
• The only references with a genuinely pre-priority, valid § 102 date that are technically on-point are US 5,809,060 (1994-02-17), US 4,707,839 (1987), US 5,103,459 (1990), US 5,357,454 (1991), and US 5,187,675 (1993) — and none of them reaches the complementary/M>N limitation.
• Date-sensitive references to watch (prior art only if the challenged claims are limited to the 1998-04-22 CIP date rather than the 1996-07-30 date): US 6,154,504, US 5,909,462, US 6,028,728, and the Bosch complementary-code decoding family. If the 1996-07-30 priority holds for the asserted claims, these drop out entirely.
• ⚠️ Do not treat US 5,841,813 or the parent US 5,862,182 as prior art: they are by the same inventor (van Nee) and/or postdate the '958 priority; the family record lists them as priority/related applications, not as prior art.

---

D. Non-patent literature (for completeness)

The record also lists 44 Non-Patent Citations, several of which are the more serious invalidity candidates than the patent citations above — notably Golay, "Complementary Series," IRE Trans. Info. Theory, Apr. 1961, C. Tseng et al., "Complementary Sets of Sequences," IEEE Trans. IT, Sep. 1972, Frank, "Polyphase Complementary Codes," IEEE Trans. IT, Nov. 1980, Wilkinson & Jones, "Minimisation of the Peak-to-Mean Envelope Power Ratio of Multicarrier Transmission Schemes by Block Coding" (VTC 1995), and the Harris "Proposed 802.11 High Rate PHY Technique — High Rate Data Modulation" (IEEE P802.11-97/144, Nov. 1997) submission. Those are § 102(b) printed publications that teach complementary codes (Golay/Tseng/Frank) and cyclic/bi-orthogonal high-rate modulation (Wilkinson, Harris). They still do not, individually, disclose the M>N, N-chip, complementary-derived code set in a single-carrier I/Q M-ary keying system — so they support § 103 combinations rather than § 102 anticipation — but they are where an invalidity attack would actually be built. I have not been able to verify every NPL citation's full contents within this session.

---

Confidence and limits (per your rules)

• High confidence: the identity of the citing patent (US 6,452,958 B1), the citation inventory itself, and each reference's number/title/assignee/filing-and-publication dates (taken from the authoritative full-text record).
• Medium confidence: the brief descriptions of references whose full text I did not open individually (I relied on the record's abstract/title text).
• Explicitly uncertain / could not confirm: (a) a direct USPTO Patent Center / PAIR confirmation of the citation list and any examiner's § 102/§ 103 rejections actually applied during prosecution; (b) the exact effective priority date the claims ultimately receive (1996-07-30 vs. 1998-04-22 CIP), which flips the prior-art status of the 1996–1998 references. Both require a direct USPTO or PACER/PAIR query that I was unable to run.

If you want, I can next build a claim-by-claim § 103 combination matrix (e.g., US 5,809,060 in view of Golay/Tseng complementary codes) — that is where these references would actually be deployed, since § 102 anticipation is not supportable on this record.

I'll ground a few key references before building the § 103 matrix.

Obviousness Analysis — U.S. Patent No. 6,452,958 B1 under 35 U.S.C. § 103

Patent: US 6,452,958 B1 ("Digital modulation system using extended code set"), van Nee, filed 1998‑04‑22, issued 2002‑09‑17, expired 2016‑07‑30. Inventor: Richard D. J. van Nee. Assignee of record: Agere Systems LLC.

Governing law: Because the application was filed before 2013‑03‑16, pre‑AIA § 103 applies. Obviousness is assessed under Graham v. John Deere (scope/content of the prior art; differences; level of ordinary skill; secondary considerations), with the expanded KSR Int'l v. Teleflex rationales. The existing section already concluded, correctly, that no cited reference anticipates any independent claim. This section therefore builds entirely on the § 103 layer: which combinations reach the claims, and why a POSITA would make them.

Note on the task date: the instructions in this task state April 26, 2026; the environment header states 2026‑09‑27. As in the earlier sections, the discrepancy is immaterial to the analysis below. Because the patent expired 2016‑07‑30, this is a retrospective validity analysis (useful for a § 286 damages‑window defense, an opinion of counsel, or expert work), not a live‑injunction matter.

---

1. Threshold issues that control the entire § 103 analysis

1.1 Effective priority date per claim (dispositive)

This is a CIP of Ser. No. 09/057,310 (filed 1998‑04‑08), itself a CIP of Ser. No. 08/688,574 (filed 1996‑07‑30, issued as US 5,862,182). Google Patents lists priority as 1996‑07‑30, but the claims get only the filing date of the application that provides § 112 written‑description support for their subject matter.

• If the "extended code set / complementary‑code‑derived" subject matter is supported by the 1996 parent (whose title is "OFDM digital communications system using complementary codes" — i.e., complementary codes were already in the disclosure), the claims get 1996‑07‑30.
• If that subject matter is new matter added in the 1998 CIPs, the claims get 1998‑04‑22.

This flips the status of four references:

Practical consequence: the 1996 date is the applicant‑favorable case, and the strongest § 103 combination does not need any of these references. I build the primary case on art that is prior art under either date (US 5,809,060, US 4,707,839, US 5,103,459, US 5,357,454, US 5,187,675, US 5,353,352, and the 1961/1972/1978/1980 complementary‑code literature).

1.2 ⚠️ Correction to the Prior Art section — § 103(c) disqualifies two of the "closest" references

The existing Prior Art section correctly notes that US 5,841,813 (van Nee, Lucent) and US 5,909,462 (Lucent) are § 102(e) art only against an 1998‑dated claim. It does not note the consequence: pre‑AIA § 103(c)(1) disqualifies subject matter that was "owned by the same person or subject to an obligation of assignment to the same person" at the time the invention was made, when that subject matter is available only as § 102(e)/(f)/(g) art. Both references are Lucent‑assigned, and the '958 was assigned by van Nee to Lucent Technologies Inc. (recorded 1998‑09‑08). If the claims get 1998 dates, '813 and '462 are disqualified as § 103 prior art. (They remain usable as generic background on the technology, but not as the source of a claim element.)

This is important for the defensive side: the seemingly most dangerous reference — van Nee's own complementary‑code work in US 5,841,813 — is the weakest § 103 weapon, not the strongest. The primary attack must come from third‑party art.

1.3 ⚠️ Correction to the Prior Art section — the "Wilkinson & Jones (VTC 1995)" citation

The Prior Art section's NPL paragraph lists "Wilkinson & Jones, 'Minimisation of the Peak‑to‑Mean Envelope Power Ratio of Multicarrier Transmission Schemes by Block Coding' (VTC 1995)." That document does not appear in the '958 Non‑Patent Citations list in the patent text. The list contains Golay (1961), C. Tseng (1972), R. Sivaswamy (1978), Frank (1980), L. Bomer (1990), A. Gavish (1994), J. Davis (1999), H. Chen (2001), the Harris 802.11 submissions, and others — but not Wilkinson & Jones. I flag this as a likely error in the prior section; it does not change the substance, because Golay, Tseng, Sivaswamy and Frank are all present and are the better references anyway.

1.4 ⚠️ Partial refinement of the Prior Art section's central conclusion

The Prior Art section states US 5,809,060 does not anticipate because its set is "orthogonal with M = N." That is right as to the orthogonal Walsh set standing alone, but it understates the reference. Under the patentee's own counting convention, the M > N limitation is closer than the previous section allows — see § 3.4 below ("the counting‑convention argument"). I flag this as a refinement, not a reversal: '060 still does not anticipate, because it lacks the complementary‑code element. But it very substantially strengthens the § 103 case on the M > N element, and the earlier section's flat "M = N" framing should not be relied on at face value.

1.5 POSITA definition

A person of ordinary skill in the art as of 1996–1998 would be: a B.S./M.S.‑level electrical engineer with 2–4 years of experience in spread‑spectrum or digital wireless modem design (DSSS, CDMA, M‑ary keying, correlator/RAKE receivers), familiar with the IEEE 802.11 DSSS PHY, the FCC Part 15.247 processing‑gain requirement for the 2.4 GHz ISM band, Walsh/Hadamard orthogonal signalling, matched‑filter/FFT correlator banks, and the classic code literature (Barker, Gold, Kasami, Golay complementary sequences). A POSITA would be able to read Golay (1961) or Tseng (1972) and implement a complementary code set; that literature is squarely within the field of spread‑spectrum signalling.

1.6 The "invention as a whole," and the patent's own admissions

The specification admits a great deal that narrows the gap:

• The MOK system of FIG. 1 (8‑chip Walsh codes with a cover sequence, I/Q, polarity bits, 8 bits/symbol) is described as existing technology.
• The FCC's processing gain ≥ 10 requirement is described as driving chip length to N ≥ 10, and the 11‑chip Barker code is described as the incumbent 802.11 DSSS choice.
• The extended set is admittedly non‑orthogonal ("The extended code set is not orthogonal, so a non‑zero cross‑correlation value results").
• The code‑derivation technique is expressly attributed to a printed reference: "an extended code set is given below in Table 1 which derived using complementary Barker codes. Complementary Barker codes are discussed in Robert L. Frank, 'Polyphase Complementary Codes,' IEEE Transactions on Information Theory, Vol. IT‑26, No. 6, November 1980, pp. 641‑647."
• The 10 Mbps / 32‑code result is presented as an arithmetic consequence: "In this example, there are 16 codes, which can be inverted to get 32 codes."

Cost of these admissions: the patent has conceded the field, the problem, the FCC constraint, the non‑orthogonality trade‑off, the counting convention (16 codes × 2 polarities = 32), and the source of the complementary codes. What is left as the asserted point of novelty is the combination — which is precisely the kind of combination KSR addresses.

---

2. The references, mapped to the claim elements

2.1 Primary reference: US 5,809,060 (Cafarella & Fischer; Micrilor; filed 1994‑02‑17 via App. 08/369,778, CIP of 08/198,138 filed 1994‑02‑17; granted 1998‑09‑15)

This is the single most important reference, and the examiner did not cite it. It appears only in the family's "Family Cites Families" block. Its disclosure, per the patent text itself and the record:

Note on sourcing: the last two quotations are from US 6,473,449 B1, which the record describes as "Continuation of application No. 09/048,651… which is a continuation of application No. 08/369,778, filed on Dec. 30, 1994, now Pat. No. 5,809,060." A continuation may not add new matter, so this text characterizes the '060 disclosure as of its 1994 filing. I flag the derivation rather than asserting it as verbatim '060 column text — a petitioner should verify against the '060 printed specification directly.

Why this matters: '060 supplies element‑for‑element the architecture of claims 1, 2, 4, 6, 13, 16, 17, 18, 22, 23, 24, 25, 38, 39, 44, 46 — everything except (i) "M > N" and (ii) "derived from a complementary code."

2.2 The complementary‑code references (all § 102(b) printed publications)

All are § 102(b) printed publications more than one year before both candidate priority dates. None is commonly owned with the '958. All are in the patent's own NPL citation list — i.e., the applicant's own IDS put the complementary‑code art of record, which removes any "the reference is non‑analogous / was not considered" defense.

2.3 Architecture and building‑block references (all § 102(b))

2.4 802.11 working‑group submissions — handling with care

The Harris submissions cited in the '958 IDS are contemporaneous evidence of the state of the art:

• IEEE P802.11‑97/144 (Carl Andren, Harris Semiconductor, Nov. 1997): shows a table with "4 bits encoded to 16 code words / 8 chips / 1.375 Mbps" and "8 bits encoded to 2×16 code words / 8 chips," i.e., M‑ary bi‑orthogonal keying at 5.5 and 11 Mbps; and — critically — states the design requires a cover sequence: "Requires a cover sequence to avoid the Wal0 CW modulation (modified Walsh Functions)." https://grouper.ieee.org/groups/802/11/Documents/DocumentArchives/1997_docs/71442.pdf
• IEEE P802.11‑97/86 (Sept. 1997) and P802.11‑98/46 (Jan. 1998): "M‑ary orthogonal keying has been known for many decades… the spread function is picked from a set of M orthogonal vectors by the data word. Since the I and Q channels can be considered independent when coherently processed, both can be modulated this way. That allows us to pack 8 bits into each symbol… It is available for 8 and 16 chip vectors and has true orthogonality." https://grouper.ieee.org/groups/802/11/Documents/DocumentArchives/1997_docs/70863.pdf

⚠️ § 102 caveat, stated precisely: the September/November 1997 and January 1998 submissions are after 1996‑07‑30. If the claims hold the 1996 date they are not prior art (they cannot be used as the source of a claim element). They remain usable as evidence of the level of ordinary skill and of industry motivation as of the relevant time, and they become § 102(a) art if the claims fall back to 1998‑04‑22. I would plead them as background/motivation, not as § 102 art, and would not build a ground on them alone.

---

3. The § 103 combinations

3.1 Ground A — the primary ground on the independent claims

Cafarella (US 5,809,060) + Golay (1961), optionally + Tseng (1972), Sivaswamy (1978), Frank (1980) and/or Bomer (1990).

The combination is a two‑reference substitution. '060 teaches "build a modem that groups bits, picks one waveform from a multi‑waveform set, phase‑modulates, and recovers it by correlating and taking the largest magnitude." Golay/Tseng teach "use this class of waveform sets, because their summed autocorrelation sidelobes vanish, which is what you want on a multipath channel." Element‑by‑element, the combination is complete for claims 1, 4, 6, 13, 16, 17, 18, 22, 23, 24, 25, 38, 39, 44, 46 and their simple dependents.

3.2 Ground B — the "ABAB′" claims (10, 11, 19, 20, 29, 30, 35, 36, 42, 47, 48)

Same as Ground A, with Golay (1961) supplying the recursive construction of longer complementary series from shorter ones (interchange, reversal, conjugation, and the concatenation/matrix constructions that generate the ABAB′ form), plus Sivaswamy (1978) and Frank (1980) on complementary constructions. The claimed A={11}, B={10} → ABAB′={11101101} in claims 11, 30, 36, 48 is a trivial 8‑element instance of a published construction, requiring no more than routine verification by a POSITA. This is the weakest claim family (and, as flagged earlier, claim 20 recites a nine‑element {111011101} for the same A and B, an internal inconsistency that muddies the claim's scope — useful to a challenger but not itself a § 103 ground).

3.3 Ground C — the "zero‑sum autocorrelation" claims (12, 21, 31, 37, 43, 49)

Same as Ground A, with Tseng (1972) as the primary complementary‑code reference. The claim language — "for shifts in the complementary code, the autocorrelations of the complementary codes sum to zero except for the main peak at zero shift" — is the textbook definition of a complementary set of sequences, and Tseng's paper is literally titled "Complementary Sets of Sequences." Combined with '060's architecture and the M > N point (§ 3.4), these claims are highly vulnerable. Note the drafting irony: the applicant chose to define the complementary code by its mathematical property in these claims, and that property is 1961/1972 public knowledge.

3.4 The M > N element — the crux, and the counting‑convention argument ⚠️

Every independent claim contains M > N. Nothing in the four‑or‑eight‑reference set says those two characters. Three independent § 103 routes reach it:

(a) Design choice / obvious degree of expansion. Given a length‑N code family, the number of bits/symbol is bounded by log₂M. Increasing M is the only way to increase throughput at fixed N and fixed chip rate. '060's specification teaches the relationship explicitly ("an alphabet of order M = 2^N is required to represent the M = 2^N possible unique binary symbol sequences of N symbols each"; "the equivalent binary data rate is the symbol rate S multiplied by the number of bits per symbol N"). Choosing M > N is the predictable use of a known work for its known purpose, in a design space where the whole point is to push M up. KSR; MPEP 2144.04 (design choice).

(b) The counting‑convention point — flagged as a refinement of the Prior Art section. '060 claims and discloses an orthogonal set plus the antipodal (inverted) set — bi‑orthogonal signalling — and, per '454, "five bits are conveyed by transmitting one of sixteen codewords and their sixteen complements, for a total of thirty‑two codewords." '958 § 2 and the E.D. Tex. construction of "code" as "a sequence of chips" mean that a complemented codeword is still a "code." Counting the full transmitted alphabet, '060/'454 disclose M = 16 codes of N = 8 chips (or 32 of 16) — i.e., M > N. The '958 patentee counts the same way for its own benefit ("there are 16 codes, which can be inverted to get 32 codes") but counts differently when distinguishing the prior art. A POSITA would not treat "8 Walsh codes + 8 inversions" and "16 codes of length 11" as different categories of thing. Whether this rises to anticipation is contestable; whether it renders M > N obvious is not close.

(c) Cyclic shifting generates a large set from a small base. US 4,707,839 (CCSK) and the '958's own admitted Table 1 methodology ("2 codes, which are cyclically shifted… to get a total of 16 different codes") show that a base code can be multiplied into a code set by cyclically shifting it. Combining that known generator with a Golay/Tseng complementary base code is a known technique applied to a known structure, with the predictable result that the shifted set retains the base code's low periodic autocorrelation — cyclic shifts do not change a code's periodic autocorrelation function. That is exactly the "predictable variation" rationale.

3.5 Ground D — fallback‑mode / rate‑switching claims

Claims 3 (six bits + two extra), the 8‑PSK/QPSK alternatives referenced in the description, and the "same code on I and Q" receiver of FIG. 10. Supplied by '060 (M‑ary PSK, "coherent and noncoherent M‑ary phase shift keying, combined with orthogonal signalling within a single symbol; and differentially encoded coherent phase keying across two symbols, with orthogonal signalling within a symbol") plus '103,459 (Walsh selection + QPSK). Claim 3's "six and two" arithmetic does not match the FIG. 4 embodiment (4 code + 2 polarity = 6) and is ambiguous — a § 112(b) issue, and for § 103 it inherits claim 2's fate.

3.6 Ground E — look‑up‑table, sidelobe‑magnitude and scrambling dependents

• Claims 8, 27, 33 (LUT storage): US 4,529,963 ("code expansion generator") plus routine implementation choice. Storing a code set in a ROM/LUT is not a patentable distinction.
• Claims 9, 15, 28, 34, 41, 45 (sidelobes ≤ ½N): a fortiori from a complementary code set, whose summed sidelobes are zero (Golay/Tseng); individual members of the '958's own preferred Table 1 set exhibit max autocorrelation 2 against N = 11, i.e. ≪ N/2. Result‑effective variable with a defined, published relationship.
• Claims 5, 26, 32, 40 (scrambling): US 5,353,352 and US 5,357,454 (modulo‑2 scrambling of block codes — "It will be appreciated that modulo‑two addition of a scrambling code corresponds, in a Galois field, to applying an axis rotation"); the '958 spec admits the scrambler is the IEEE 802.11 standard 1‑bit/127‑chip scrambler. Admitted prior art.
• Claims 7, 14 (complementary code length 2^X): Golay (1961) — the classical binary complementary constructions produce lengths 2, 4, 8, 16, 32…

---

4. Claim‑by‑claim § 103 matrix

Strength key: ●●●●● = very likely invalid · ●●●●○ = likely invalid · ●●●○○ = contestable · (claims are grouped where the analysis is identical)

Aggregate read: the independent claims that recite the generic "derived from a complementary code" limitation (1, 13, 22, 32, 38, 44) are strong § 103 targets; the zero‑sum‑property family (12, 21, 31, 37, 43, 49) is the strongest; the ABAB′ family (10, 11, 19, 20, 29, 30, 35, 36, 42, 47, 48) is slightly weaker only because the specific construction must be matched to Golay's 1961 presentation — a proof problem, not a substance problem.

---

5. Motivation to combine — the KSR / MPEP 2143 rationales

Each rationale below is independently sufficient; together they are overwhelming.

R1 — Same field, same problem, same solution family (MPEP 2143.01(I); KSR). Both '060 and the '958 address a high‑data‑rate wireless LAN over a multipath channel in the 2.4 GHz ISM band. Both identify multipath/delay spread as the barrier. Both solve it by lengthening the symbol (higher‑order alphabet + processing gain) rather than by narrowing the symbol. There is no field boundary to cross.

R2 — The FCC constraint supplied the design target. The patent itself states the target: a processing gain ≥ 10, hence N ≥ 10 chips. Given N ≥ 10 and the need for > 8 bits/symbol to hit 11 Mbps in the 802.11 channel spacing, M must exceed N. The prior art thereby supplies the incentive, and the '958 supplies only the arithmetic. In re Kemps / KSR ("design incentives and other market forces").

R3 — The complementary‑code literature existed precisely to solve the claimed sub‑problem. Golay's property is "autocorrelation sidelobes that vanish," and Tseng extended it to sets. The '958's claimed function — "autocorrelation sidelobes suitable for multipath environments" — is the stated purpose of that entire literature. MPEP 2143.02 (a reference teaching the claimed function/desired property). The applicant's own specification cites Frank 1980 as the source of the code set in Table 1, which is a written admission that the codes came from the art.

R4 — Known technique applied to a known structure, predictable result (MPEP 2144.04; KSR). CCSK ('839) teaches generating a code set by cyclically shifting a base code; the '958's own preferred Table 1 does exactly that. Because cyclic shifting preserves periodic autocorrelation, applying the shift generator to a complementary base code yields a larger set with the low sidelobes retained — the result the artisan would predict, with no unexpected property to be discovered.

R5 — Bi‑orthogonal bit‑packing is standard, and the arithmetic is identical ('454). The '958's headline number — 16 codes + 16 inversions = 32 codewords → 5 bits per branch → 10 bits/symbol — is stated verbatim in structure in US 5,357,454 ("one of sixteen codewords and their sixteen complements, for a total of thirty‑two codewords… five bits"). Compressing 8+ bits/symbol was the contemporaneous industry objective (Harris's 802.11 submissions showing "8 bits encoded to 2×16 code words").

R6 — Modifying a Walsh/MOK set with a cover sequence was industry practice, not invention. Harris's own submission states the technique is required ("Requires a cover sequence to avoid the Wal0 CW modulation (modified Walsh Functions)"), and the '958's own IDS/co‑pending application (Ser. No. 09/057,310 → US 6,404,732) is directed to precisely that. The step from "cover‑sequence‑modified orthogonal set" to "set derived from a complementary code" is a substitution of one known sidelobe‑shaping tool for another with the same function.

R7 — Receiver elements were off‑the‑shelf and on the record. FWT processor ('454), maximum search circuit ('675), RAKE/diversity ('390, '307), subtractive demodulation ('619, '919). All were cited by the examiner. An obviousness rejection may combine references that the examiner cited with references that the applicant cited; the applicant's own IDS put Golay, Tseng, Frank, Sivaswamy, Bomer, and Gavish on the record, which forecloses a "not considered" defense.

R8 — No teaching away. The only candidate teaching‑away argument is that the prior art prized orthogonality ('459, '352, '060's Walsh set) whereas the '958 set is admittedly non‑orthogonal. That fails on three grounds: (i) '060 expressly contemplates non‑orthogonal higher‑order alphabets ("the use of a non‑orthogonal high‑order signalling alphabet results in acceptable performance"); (ii) CCSK ('839) is itself a non‑orthogonal code alphabet and was long‑accepted practice; (iii) the '958 specification admits non‑orthogonality and treats the resulting degradation as tolerable — you cannot teach away from a trade‑off the patent itself accepts. In re Fulton; MPEP 2145 (teaching away must be such that a POSITA would be led in a divergent direction).

R9 — The claim scope is unbounded where the alleged invention is specific. No claim recites 11 chips, 16 codes, 10 Mbps, or a chip rate of 11 MHz. The only numerical limitations are M > N (independent), 2^X code length (claim 7), and ≤ ½N sidelobes (claims 9 et al.). Where the patentee's asserted advance is the specific 16×11 / 10 Mbps configuration but the claims recite only the generic relationship, the claims are commensurately broad — and broad, generic claims are the easiest to invalidate. (This is the mirror image of the 2008 Markman posture: the broad constructions Agere obtained for infringement purposes are a liability on validity.)

---

6. Rebuttal: what survives, and what a patentee would argue

Procedural note: because no IPR/PGR/CBM was ever filed against the '958 (per the PTAB section above), no § 315(e)(2) or § 325(e)(2) estoppel attaches to anyone, and the § 315(e)(1) serial‑petition bar is not a risk. If this analysis were driving a live challenge, the prior‑art universe above would be fully available. Conversely, the patent expired 2016‑07‑30, so a post‑grant challenge buys a Board statement rather than injunctive relief.

---

7. Bottom line

1. No anticipation. The prior section's conclusion holds: no single reference discloses M > N and "derived from a complementary code."
2. The claims are nevertheless exposed under § 103, and the exposure is concentrated in the generic complementary‑code families (claims 1, 12, 13, 21, 22, 31, 37, 44, 49 and their dependents) rather than in the ABAB′ family.
3. The primary combination is Cafarella (US 5,809,060) + Golay (1961), optionally adding Tseng (1972), Sivaswamy (1978), Frank (1980) for the complementary‑set property, and '454 / '675 / '459 / '839 / '529,963 / '352 for the receiver, generator and scrambling elements. The single most important fact in this analysis is that the examiner never cited US 5,809,060 — the reference that supplies nearly every claim element except the complementary‑code and M > N limitations, and which is available as prior art under either candidate priority date.
4. M > N is the weakest link in the patent, not the strongest, because (i) it is an unbounded design choice over an admitted goal, (ii) the patentee's own counting convention makes bi‑orthogonal prior‑art sets satisfy it, and (iii) no claim recites the specific 16×11 configuration that gives the patent its story.
5. The ≤ ½N sidelobe claims (9, 15, 28, 34, 41, 45) and the zero‑sum claims (12, 21, 31, 37, 43, 49) are the easiest to invalidate, because they recite by definition the known property of a 1961/1972 published code class.

Recommended next step if this were a live matter: (a) resolve the effective priority date claim‑by‑claim by comparing each independent claim's limitations to the 08/688,574 and 09/057,310 disclosures — this determines whether the 1996‑to‑1998 references are available as § 102 art; (b) pull the printed specifications of US 5,809,060 and US 5,357,454 (rather than relying on the continuation's text) to lock down the quotes; (c) obtain Golay 1961 and Tseng 1972 in full to fix the exact ABAB′/recursive construction passages for the claim 10/11 family; and (d) confirm the negative PTAB finding in PTAB E2E.

---

Confidence and limits

• High confidence: the claim language and element decomposition (taken verbatim from the patent text in my context); the identity/dates/assignees of the prior‑art references; the existence and content of the '060, '454, '675 and Harris disclosures I quote (each corroborated by at least one independent source this session); the § 103 legal framework and the § 103(c) common‑ownership point; that no single reference anticipates.
• Medium confidence: the precise wording of the '060 specification passages I quote via US 6,473,449 (a continuation of the '060 application — no new matter permitted, but the quotes should be verified against the '060 printed specification); Golay's inclusion of the specific ABAB′ construction form (I could confirm Golay's recursive/invariance constructions and the "sum of autocorrelations vanishes at nonzero delays" property, but not the literal ABAB′ nomenclature in the 1961 paper); that a POSITA would read Tseng 1972 as teaching "code sets" satisfying the claim 12/21/31/37/43/49 definition (highly likely from the title and 1972 abstract, but I did not obtain the full paper).
• Unresolved / flag for verification: the effective priority date of the asserted claims (1996‑07‑30 vs. 1998‑04‑22), which is determinative; whether US 5,809,060's full printed specification contains the non‑orthogonal‑alphabet passage; and the accuracy of the earlier section's "Wilkinson & Jones (VTC 1995)" NPL citation, which I could not locate in the '958 citation record and believe to be erroneous.
• Explicit non-findings: I found no evidence that any of these combinations was previously raised — in prosecution, at the ITC, in the E.D. Tex. case, or at the PTAB. The examiner's failure to cite US 5,809,060 and Tseng is the central substantive observation of this analysis.