US 6801579

I'll search for litigation/docket information and verify details for this specific patent number.

Let me check for any Federal Circuit docket activity in 2026 and USPTO records specific to this patent.

US Patent 6,801,579 B1 — Analyst Summary

Sourcing note: The patent's full text and bibliographic data were taken from the Google Patents record (https://patents.google.com/patent/[US6801579](/patent/US6801579)/en). I ran targeted searches for any 2026 CAFC docket or USPTO proceeding specifically naming 6801579. I found no authoritative record of a 2026 Federal Circuit appeal or other 2026 litigation involving this exact patent number. The only third-party hits were (a) reference to "6801579" in unrelated PTAB IPR exhibits (as citation material) and (b) Brazilian design registration "DI 6801579-8," which is a different, unrelated identifier (a footwear design, Wolverine World Wide) and must not be conflated with the US patent. Treat absence of 2026 docket results as "not found," not as proof of non-existence.

Bibliographic Data

Abstract (paraphrased)

A method for wireless transmission from an array of two or more antennas in which each transmitted signal is selected from a constellation of unitary space-time signals. Each signal is a unitary matrix whose columns represent antennas, rows represent time intervals, and elements represent complex amplitudes. In embodiments, the constellation matrices form a non-Abelian group with positive diversity product, or a coset of such a group; alternatively the constellation is a subset whose multiplicative closure is such a group; still alternatively the constellation is an extension adding elements outside the group.

Plain-Language Overview of Independent Claims

Transmitter-side claims (1, 8, 9, 12): each recites selecting a signal from a constellation of unitary-matrix signals, then transmitting from two or more antennas across two or more time intervals, with amplitudes proportional to the matrix elements (rows = time intervals, columns = antennas; claim 12 limits to exactly two antennas).

• Claim 1 — Selection made from a constellation whose matrices are any one of: (I) a non-Abelian group, other than a quaternion group, having a positive diversity product; (II) a subset of such a non-Abelian group whose multiplicative closure is itself such a group; (III) a coset of (I) formed by left, right, or both multiplications; or (IV) a coset of (II).
• Claim 8 — Selection from a constellation at least half of which corresponds to matrices expressible as products A^l B^k (or cosets). A is an n×n diagonal matrix of integer powers of an m-th root of unity; B is an n×n shift-type matrix with a root-of-unity entry (e^(2πi/s)) in the bottom-left; with s, m positive integers, n ≥ 2, l < m, and k < min(s, n). (This tracks the specification's generalization of group G_{m,r} to the non-group set S_{m,s}.)
• Claim 9 — Selection from a constellation at least half of which corresponds to matrices expressible as pairwise products of elements of two unitary representations of finite groups G_A and G_B, or cosets thereof. (Dependent claims 10–11 add that the groups are fixed-point-free, and that G_A = G_B as equivalent reducible representations of an odd-order group.)
• Claim 12 — Two-antenna method using a "Hamiltonian" constellation: at least half the signals are 2×2 unitary matrices of the form [[x, −y*],[y, x*]], with |x|² + |y|² = 1, and the set includes at least one matrix where ‖x| − |y‖ > 0 (i.e., |x| ≠ |y|). Dependent claim 13 requires that difference to be at least 0.1.

Receiver-side claims (14, 21, 22, 25): mirror claims 1, 8, 9, and 12 respectively, but recite the receive method — receiving the multi-antenna, multi-interval signal and comparing it against the constellation to identify which selected signal was sent (i.e., the detection/decoding counterpart). Corresponding dependents: 15–20, 23–24, 26.

Notable Literal-Reading Items (no auto-correction applied)

The claims contain drafting irregularities worth flagging if analyzing claim scope verbatim:

• Claim 1 (and claim 14) recite "guaternion group" in element (II) while element (I) uses "quaternion group," and both use "quaternion group" in the exclusions — i.e., the exclusion from non-Abelian groups applies to quaternion groups.
• Claim 12 lists element "(m)" rather than "(III)," and has a reversed-bracket typo in the matrix: [x -y* / y x*. ].
• Claim 8 describes B as a "diagonal matrix" although the depicted matrix is a cyclic-shift (companion) matrix; claim 8 also states "i = −1" whereas claim 21 states "i² = −1."

Uncertainty / Limitations

• Assignee history is per recorded assignments (Lucent → Alcatel-Lucent → WSOU Investments, with Nokia of America listed); I did not independently verify current chain of title beyond the record.
• The patent is expired (adjusted expiration 2022-12-21), which materially limits the practical significance of any 2026 enforcement activity.
• I could not confirm any CAFC 2026 docket naming 6801579; the docket results I retrieved concerned other patents and parties. If a current 2026 docket exists, it was not surfaced by these searches and would need direct PACER/CAFC or Docket Navigator verification.

I'll run targeted searches for litigation naming this exact patent number.

Litigation Search — US Patent 6,801,579 B1

Bottom line: I found no litigation — district court, ITC, or Federal Circuit — in which US 6,801,579 is identified as an asserted patent or patent-in-suit. Everything the number surfaced in was either (a) prior-art/technical citation inside other parties' proceedings, or (b) an unrelated foreign identifier.

I was cut off by a tool-step limit partway through the search sequence, so treat this as a strong preliminary negative rather than an exhaustive PACER sweep (see "Limits of this search" below).

---

1. Cases found asserting US 6,801,579

No case in any of my result sets lists 6,801,579 among its patents-in-suit.

---

2. Look-alike results I checked and ruled out (do not conflate)

---

3. Where the number 6,801,579 does appear: other parties' PTAB filings, as prior art

The only substantive US proceedings in which the number appeared were IPR petition exhibits citing it as a printed publication, not as a challenged or asserted patent:

• USPTO PTACTS petition 1479356 — citing "Method for Wireless Communication using Unitary Space-Time Signal Constellations," #6801579, October 5, 2004 (in argument about an unrelated '607 patent). https://ptacts.uspto.gov/ptacts/public-informations/petitions/1479356/download-documents
• USPTO PTACTS petition 1554449 — same citation, in a secondary-considerations / prosecution-history discussion for an unrelated '486 patent. https://ptacts.uspto.gov/ptacts/public-informations/petitions/1554449/download-documents

This is consistent with the patent's role in the literature as a foundational reference on unitary space-time constellations (cf. the sole cited prior art, US 6,724,842 B1, Lucent, and the cited non-patent literature such as Hassibi et al., Bell Labs Technical Memorandum, Mar. 2000).

---

4. Consistency checks against the record

• Google Patents record shows no litigation section for this patent. The fetched page includes citation tables ("Cited By (19)", "Families Citing this family (5)", "Citations (1)") but no litigation/assertion entries — the Google Patents litigation module is populated from external docket data, so its absence is corroborating (though not conclusive) evidence of no recorded US assertions. Source: https://patents.google.com/patent/US6801579/en
• Expiry materially explains the absence. Status is "Expired – Lifetime," adjusted expiration 2022-12-21. Even if a case once existed, no forward-looking enforcement is possible now.
• Assignee chain relevant to any campaign: Lucent Technologies → (Alcatel-Lucent) → WSOU Investments LLC (recorded 2017-09-25), with security interests to Omega Credit Opportunities Master Fund (2017) and BP Funding Trust Series SPL-VI (2019), both released. WSOU (a Craig Etchegoyen NPE, d/b/a Brazos Licensing and Development) is the only entity in the chain that has actually run assertion campaigns — and, as noted, it did not assert this patent in the cases I could identify.

---

5. Contradictions / date issues to flag explicitly

1. Internal date inconsistency in the intake materials. Your task header says current date April 26, 2026; the operative system date is 2026-09-28; the Google Patents fetch is stamped 2026-03-26. These are not reconcilable as stated. Note that my search results did return 2026 docket activity (e.g., WSOU v. Google, 6:20-cv-00580, with entries through mid-2026, and In re WSOU Investments, Fed. Cir. 2026) — but none of it concerns 6,801,579.
2. Prior section's "quaternion group" reading stands and is relevant here only tangentially: claim 1 element (II) recites "guaternion group" while (I) recites "quaternion group." I did not auto-correct this. A constellation literally branded a quaternion group falls inside the express exclusion in claim 1 — e.g., the specification's D₄,₁,₋₁ / "quaternion group Q₆" examples — which is a scope-limiting oddity worth noting in any freedom-to-operate or claim-charting exercise, though it is moot given expiry.
3. No 2026 CAFC appeal naming 6,801,579. This confirms (rather than contradicts) the prior section's finding. The Fed. Cir. docket for In re WSOU Investments LLC (No. 25-1153, https://www.courtlistener.com/docket/74780398/in-re-wsou-investments-llc/) is a WSOU-wide matter and I found no indication it involves this patent.

---

6. Limits of this search — and how to close the gap

• My searches were terminated by a tool-step cap before I could query Unified Patents' caselist (https://portal.unifiedpatents.com/litigation/caselist), RPX Insight, Docket Navigator, or PACER/ITC EDIS directly. The negative result is based on CourtListener/RECAP text, PTACTS documents, and general web indexing — not a complete docket-system query.
• ITC Section 337 investigations (which the DED "Complainants/Respondents" document hints at) are not fully indexed in CourtListener; a direct EDIS search is the right check.

Recommended confirmatory steps before relying on this negative:

1. Search Unified Patents caselist and RPX Insight for patent number 6801579 (exact-match).
2. PACER/Recap full-text search restricted to "6,801,579" across all district courts; separately run "6,801,579" in ITC EDIS.
3. Pull the litigation history tab on Lex Machina / Docket Navigator for the patent number, which would also capture any Lucent-era assertion (pre-2017) that predates modern free dockets.

If a case does exist, it is most likely an older Lucent/Alcatel-Lucent assertion (2004–2012) that is not surfaced in free sources, or alternatively the unexplained "'680 patent" shorthand in D. Del. docket 40334 — which I was unable to tie to this patent number and which you should not treat as confirmed.

I'll verify the structured block against live sources before concluding.

Proceedings overview

Total AIA trial proceedings on US 6,801,579: 0. The canonical structured source (USPTO Open Data Portal "PTAB proceedings on file" block) returns no IPR, PGR, or CBM trial for this patent, and my independent web searches found none either — so the breakdown is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution-denied.

Defensive posture: this is not a "hardened by IPR" patent and not a "claims already canceled" patent — it is an untested patent. All 26 claims stand as issued, unexamined by any Article I tribunal, and the patent went to its grave (adjusted expiration 2022-12-21) without ever having been challenged at the Board. A defendant today gets neither the gift of canceled claims nor the comfort of a Federal Circuit blessing; it inherits a blank slate plus the legal reality that the patent is expired and any damages window is narrow.

---

No proceeding numbers to report

Because the count is zero, I am deliberately not creating any proceeding headers. Per the task constraints, I will not invent proceeding numbers, petitioners, panels, or FWD dispositions. Anyone claiming "the '579 patent survived two IPRs" or "claims 1–5 were canceled in IPR20xx-xxxxx" would be fabricating; neither statement is supportable from any source I could reach.

---

Near-misses and lookalikes (things that look like PTAB hits but are not)

These are the artifacts my searches surfaced. Each was checked and excluded; I flag them because they are the likeliest source of a false positive in a docket-scrub.

I also searched for ex parte reexamination, reissue, and district-court invalidity challenges specifically naming 6,801,579 and found no authoritative record. The assignee-side litigation context I did surface is WSOU Investments' 2020–2021 W.D. Tex. campaign before Judge Albright (Canon, Dell/EMC, Netgear, OnePlus, TP-Link, Google, Microsoft, etc.), which asserted various Lucent/Alcatel-Lucent patents — the surfaced complaints assert the '714, '346, '536 and similar patents. I could not confirm that 6,801,579 itself was asserted in any of those complaints. Treat "WSOU asserted '579" as unverified.

---

Strategic summary

Claim status — CANCELED: none. SUSTAINED: none. UNTESTED: all 26. Claims 1 and 14 (the non-Abelian-group-including-cosets genus claims, with the quaternion-group carve-out), claims 8 and 21 (the A^l B^k / S_{m,s} generalization), claims 9 and 22 (pairwise products of two group representations), claims 12 and 25 (Hamiltonian constellations with |x| ≠ |y|), all dependents, and the entire receiver-side family remain exactly as they issued on 2004-10-05. There is no claim-level disposition to cite, because there is no FWD to cite.

Estoppel landscape — § 315(e)(2) is a non-issue, which cuts both ways. With no prior petition, no petitioner and no privies are estopped, so every prior-art ground remains fully available to a defendant: § 102/§ 103 over the on-face citation US 6,724,842 B1 (Hochwald et al., Lucent, filed 1999-07-16 — a sibling application, available as pre-AIA § 102(e) art given the 2000-03-09 priority date), the printed publications in the Non-Patent Citations list (Tarokh et al. 1998; Hughes, "Differential Spacetime Modulation," WCNC 1999; Hochwald et al., Bell Labs TM, Jan. 1999; Hassibi et al., Bell Labs TM, Mar. 2000), and — importantly — the Ramot WO 2005/074163 and other later "extension of a fixed-point-free group" art that the '579 itself anticipates. Nothing is walled off. The flip side is that the patent owner has never had its claims construed by anyone, so there is no narrow construction or prosecution disclaimer to leverage from a PTAB record.

Pattern signals. There is no pattern to read. No repeat petitioner, no defensive aggregator (no Unified Patents, no RPX, no IP Navigation-type filer) took a shot at this patent, and the patent owner never had occasion to defend a PTAB appeal to the Federal Circuit. That absence is itself the signal: a 2000-priority Lucent/Alcatel-Lucent patent on unitary space-time constellations — squarely in the 4G/5G MIMO lineage, and cited by Nokia/Rampart/NEC/New Jersey Institute of Technology follow-on work — survived its entire 18-year term without a single AIA challenge. That is consistent with low assertion value (expired, crowded art field, and the practice of asserting older video/networking Lucent patents instead), not with demonstrated invalidity-proof strength.

---

Recommended next steps

1. Do not represent that this patent has been tested. If you are a defendant and opposing counsel or a prior internal docket note refers to an IPR or FWD on 6,801,579, demand the proceeding number and pull it — based on everything reachable today, it does not exist. The AIA trial lookup at https://ptacts.uspto.gov/ptabs/ and the ODP record both come back empty.
2. Exploit the blank slate while it lasts. Because no PTAB estoppel attaches and no claim has been construed, a defendant facing a demand can front-load § 102/§ 103 theories built on US 6,724,842 B1 and the 1998–2000 Bell Labs/Tarokh/Hughes literature with no estoppel risk. Note, however, that a post-2022-12-21 expiration means the practical fight is over the residual damages window — under § 286 the lookback runs only to 2020-09-28, giving a theoretical six-year tail of about 2020-09-28 through 2022-12-21.
3. Confirm the chain of title you are dealing with. Recorded assignments run Lucent → Alcatel-Lucent → (Credit Suisse/Omega/BP Funding security interests) → WSOU Investments LLC, currently listed with Nokia of America Corp. As the prior summary already flagged, I did not independently verify current title. If a demand letter arrives citing 6,801,579, verify that the sender actually holds the right to sue on the post-2022-12-21 tail.
4. If you need a PTAB record, there isn't one to link. I cannot supply a USPTO PTAB E2E or CourtListener link for a disposition of this patent, because no such disposition exists. Links here are limited to the lookalike artifacts in the table above, which I have explicitly disclaimed.

I'll search for the assignment record details for this patent.

Let me dig deeper into the assignment chain specifics and any assertion history.

Good context on the WSOU chain. Now let me try to pin down the specific reel/frame entries for US 6801579.

Assignment & Ownership Analysis — US 6,801,579 B1

Sourcing caveat up front. The event list below is reconstructed by cross-referencing three sources: (1) the Google Patents "Legal Events" record for US 6,801,579 (authoritative, reproduced in the full text supplied to me); (2) USPTO reel/frame citations quoted inside federal court filings in the WSOU Investments litigation family (Cisco, Juniper), which are the only place I could obtain reel/frame numbers at all; and (3) secondary NPE-directory and RPX material. My search tooling does not render the Assignment Center's dynamic results grid, so I was unable to pull the correspondent-of-record field or the confirmed reel/frame for this exact patent. I have labeled inferred reel/frames as such. I have not invented any.

---

Inventors

Basis / confidence: The specification repeatedly references four co-pending applications "all commonly assigned herewith" (Ser. Nos. 09/134,297; 09/206,843; 09/356,387; 09/528,973) and a Bell Labs Technical Memorandum (Hassibi et al., March 2000) — consistent with an all-Bell-Labs inventor group. The Bell Labs Technical Journal also lists Wim Sweldens among Lucent patentees of that era.

Unusual-pattern check: I could not verify departure dates from my retrieved sources, so I flag the "all inventors gone within 12 months" pattern as not established. I will note only that this is a pure research-lab artifact (four Bell Labs mathematicians, no product engineer), which is a mild forward-indicator that no commercial product line ever embodied the claims — but the assignment record shows no inventor-originated spin-out, so nothing here suggests an early fire-sale driven by inventor attrition.

---

Original assignee

• Entity on the issued patent: Lucent Technologies Inc. (Murray Hill, NJ), per the 2000-08-22 assignment of assignors' interest recorded in the Google Legal Events record, original assignee field.
• Primary line of business: Lucent was the AT&T equipment spin-off (1996) — a major telecom network-equipment manufacturer (switching, optical, wireless infrastructure). It was an operating company, not an IP holding vehicle.
• Did they ship a product embodying the claims? Unclear / no evidence found. In 2000 unitary space-time modulation (USTM) was a Bell Labs research program; the technique later influenced MIMO/space-time coding in 3GPP, but I found no Lucent product identified as practicing the claims of the '579 patent. Treat as research-stage technology, not a shipped product.
• Current status: Lucent merged with Alcatel (France) on 2006-12-01, forming Alcatel-Lucent; Alcatel-Lucent was in turn acquired by Nokia (completed 2016). The Alcatel-Lucent US patent-holding entity (Alcatel-Lucent USA Inc.) is now inside Nokia. No bankruptcy proceeding is implicated for this chain.
• Recorded assignee history on this patent: Lucent Technologies Inc. → (Alcatel-Lucent USA Inc., merger/name change — not shown as a discrete recorded assignment in the Google Legal Events list) → Alcatel Lucent (assignor on the 2017-09-25 record) → WSOU Investments, LLC.

---

Assignment timeline

Note on reel/frame and correspondents: Unless labeled "confirmed," reel/frame values are contextual — drawn from court filings describing the same WSOU/Acatel-Lucent recording family on the same dates, not from a direct Assignment Center query on 6801579. No correspondent-of-record names were retrievable for any link; I will not guess them.

• 2000-08-22 (recorded; execution date not separately shown) — Reel not retrieved

  • Conveyance: Assignment of assignors' interest (SEE DOCUMENT FOR DETAILS)
  • Assignor: Hassibi, Babak; Hochwald, Bertrand M.; Shokrollahi, Mohammad Amin; Sweldens, Wim
  • Assignee: Lucent Technologies Inc.
  • Correspondent: not retrieved
  • Context: Original assignment of inventors' rights to the employer on filing.

• 2006-12-01 (corporate event; NOT a recorded assignment) — Reel n/a

  • Conveyance: Merger (Lucent Technologies Inc. → Alcatel-Lucent)
  • Assignor: Lucent Technologies Inc.
  • Assignee: Alcatel-Lucent (US patent-holding sub: Alcatel-Lucent USA Inc.)
  • Correspondent: n/a
  • Context: Internal reorg / merger. Gap: this merger does not appear in the patent's Legal Events list even though later events name "Alcatel-Lucent USA Inc." as an assignor — a chain-of-title hole worth verifying at the Assignment Center.

• 2013-03-07 (recorded) — Reel not retrieved

  • Conveyance: Security Interest
  • Assignor: Alcatel-Lucent USA Inc.
  • Assignee: Credit Suisse AG
  • Correspondent: not retrieved
  • Context: Securitization — collateral pledge under an Alcatel-Lucent credit facility (encumbrance, not a title transfer).

• 2014-10-09 (recorded) — Reel not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: Credit Suisse AG
  • Assignee: Alcatel-Lucent USA Inc.
  • Correspondent: not retrieved
  • Context: Lien release — encumbrance cleared.

• 2017-09-21 (recorded) — Reel not retrieved

  • Conveyance: Security Interest
  • Assignor: WSOU Investments, LLC
  • Assignee: Omega Credit Opportunities Master Fund, LP
  • Correspondent: not retrieved
  • Context: Acquisition financing / securitization — note this lien is recorded four days before the title assignment lands, i.e., pledged as part of the closing.

• 2017-09-25 (recorded) — Reel 044000/0053 (likely — see note)

  • Conveyance: Assignment of interest
  • Assignor: Alcatel Lucent
  • Assignee: WSOU Investments, LLC (11150 Santa Monica Blvd, Suite 1400, Los Angeles, CA 90025)
  • Correspondent: not retrieved
  • Context: Transfer to asserter. Executed through the Nokia/Alcatel-Lucent → Wade and Company (Ontario) Patent Purchase Agreement structure, then Wade's PPA interest → WSOU (Wade/WSOU Agreement, effective 2017-08-21).
  • Reel note: In the Cisco litigation, the Sept 25 2017 recording — comprising Amended Schedule B1 (Alcatel Lucent assignment), the Wade/WSOU Agreement, the WCFT Cayman release, and the 149-page assigned-patents exhibit — is cited as USPTO Reel 044000 Frame 0053. I could not confirm that 6801579 is listed within that exhibit. A separate Alcatel-Lucent→WSOU batch executed 2017-12-22 (229 properties, EPAS ID PAT4782264) was recorded at Reel 045085 Frame 0001; membership of '579 there is also unconfirmed.

• 2019-05-20 (recorded) — Reel not retrieved

  • Conveyance: Security Interest
  • Assignor: WSOU Investments, LLC
  • Assignee: BP Funding Trust, Series SPL-VI
  • Correspondent: not retrieved
  • Context: Refinancing / securitization — new collateral pledge.

• 2019-05-21 (recorded) — Reel not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: OCO Opportunities Master Fund, L.P. (f/k/a Omega Credit Opportunities Master Fund LP)
  • Assignee: WSOU Investments, LLC
  • Correspondent: not retrieved
  • Context: Lien release — 2017 Omega encumbrance cleared, coincident with the BP Funding Trust pledge.

• 2021-06-03 (recorded) — Reel not retrieved

  • Conveyance: Release by Secured Party
  • Assignor: Terrier SSC, LLC
  • Assignee: WSOU Investments, LLC
  • Correspondent: not retrieved
  • Context: Lien release. Gap: no corresponding grant of security interest to Terrier SSC appears in the Legal Events list — an intermediate secured party or successor whose grant record is missing from the captured data.

• 2022-12-21 — Adjusted expiration (patent term end; not an assignment) — Reel n/a — Context: Expiry. Patent is now listed Expired – Lifetime.

---

Timeline diagram

timeline
    title Ownership of US 6801579
    2000 : Inventors assign to Lucent Technologies
    2004 : Patent issued
    2006 : Lucent merges into Alcatel-Lucent
    2013 : Credit Suisse security interest
    2014 : Credit Suisse release
    2017 : Nokia entities sell portfolio to Wade
         : Wade transfers PPA to WSOU Investments
         : Alcatel Lucent assigns title to WSOU
         : Omega Credit security interest
    2019 : BP Funding Trust security interest
         : Omega release
    2021 : Terrier SSC release
    2022 : Patent expires

---

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT (strong).
Title moved from an operating company (Alcatel-Lucent/Nokia) to a licensing-only LLC. Concrete evidence: the 2017-09-25 record transfers the asset to WSOU Investments, LLC, and Cisco's Rule 12(b)(1) motion in WSOU v. Cisco, No. 6:21-cv-00128-ADA (W.D. Tex.) states flatly that "WSOU is an NPE founded by Craig Etchegoyen[;] [i]ts primary business is threatening and filing lawsuits." WSOU was formed in Delaware in July 2017 with a single-purpose address at 11150 Santa Monica Blvd, Suite 1400, Los Angeles, CA 90025 (per the recorded assignment cover sheet), and litigates under the trade name Brazos Licensing and Development. No product line exists.

2. Known asserter in the chain — PRESENT (strong).
WSOU Investments appears in the Stanford NPE Litigation Database (listed as "Non Practicing Entity," e.g., WSOU Investments, LLC v. ZTE Corp., 6:20-cv-00211, W.D. Tex.) and is a recurring subject of RPX high-frequency-plaintiff reporting (e.g., RPX's 2020-03-28 note on nine new Huawei cases). The Etchegoyen link ties the chain to the Uniloc family of asserters. This is a documented asserter, not a name-based inference.

3. Repeat correspondent across the chain — UNCLEAR / not verifiable.
I could not retrieve the correspondent-of-record for any recorded assignment in this chain. This is the single most significant evidence gap in this report; it should be closed by a direct Assignment Center query, because a recurring recording attorney across the 2017/2019/2021 WSOU-filed entries would corroborate the controlled-vehicle pattern.

4. Cascading transfers — PRESENT (moderate).
Within roughly 90 days in 2017 the asset moved through Nokia entities → Wade and Company (PPA, effective 2017-07-22; amended 2017-08-02) → WSOU (Wade/WSOU Agreement, 2017-08-21) → recorded to WSOU (2017-09-25), with the Omega security interest recorded 2017-09-21. A second WSOU-filed title recording executed 2017-12-22 (Reel 045085/0001) shows a further batch channel. The chain then carries successive collateral pledges (Omega 2017 → BP Funding Trust 2019) and releases (2019, 2021) — the classic financed-asserter structure.

5. Pre-litigation transfer — UNCLEAR.
The 2017 transfer predates the WSOU litigation wave (2020–2021) by ~3 years, which is not a within-6-months pre-suit transfer. However, I found no suit asserting 6801579 specifically, so I cannot date a "first suit naming this patent." Call it not applicable on current evidence rather than present.

6. Bankruptcy fire-sale — NOT PRESENT.
The Nokia/Alcatel-Lucent divestiture was a strategic portfolio sale (facilitated by Aqua Licensing; the same program spun assets into Provenance Asset Group LLC and Wade and Company), not a Chapter 7/11 liquidation.

7. Privateering — UNCLEAR (leaning present, but not proven).
The Nokia/Alcatel-Lucent portfolio was sold to an NPE that then sued Nokia's competitors (Huawei, ZTE, Cisco) — the privateering signature. Reinforcing evidence: the recorded Amended Schedule B1 assignment reserves "encumbrances and retained rights… in favor of Assignor and/or its assignees and licensees," i.e., a seller license-back, which is a common privateering term. I did not, however, locate an SEC filing or a revenue-share/retained-assertion right proving Nokia directs or profits from WSOU's suits, so I stop short of "present."

8. Defensive aggregator — NOT PRESENT.
The chain terminates at an active asserter (WSOU), not at RPX/AST/LOT/Unified/OIN. The patent has not been neutralized; it simply expired 2022-12-21, which now moots assertion.

---

Verdict

NPE — high confidence.

Two independent strong signals are documented, not inferred: (i) a shell-entity transfer — the 2017-09-25 recorded assignment moves the patent from operating-company ownership (Alcatel-Lucent/Nokia, via the Wade and Company PPA) to WSOU Investments, LLC, a Delaware single-purpose licensing vehicle with no products and an office-suite address; and (ii) a known asserter in the chain — WSOU is listed as a non-practicing entity in the Stanford NPE database and is a documented high-volume filer (RPX; WSOU v. Cisco, 6:21-cv-00128-ADA, where Cisco states WSOU's business is "threatening and filing lawsuits"). The 2019/2021 security-interest → release entries (BP Funding Trust, Omega/OCO, Terrier SSC) further mark the asset as securitized collateral of a monetization vehicle. The one pattern I could not substantiate — a repeat correspondent of record — is an evidence gap, not a negative finding, and should be closed by querying the Assignment Center directly.

Verify at: https://assignmentcenter.uspto.gov/ (or https://assignment.uspto.gov/patent/index.html) — search patent number 6801579; confirm the reel/frame for the 2017-09-25 entry and capture the correspondent-of-record for each link.

Key limitation to carry forward: this patent expired 2022-12-21 ("Expired – Lifetime"), so the NPE classification describes the historical ownership structure, not a live assertion threat.

I'll verify the citation record for US 6,801,579 and research each cited reference before building the prior-art analysis.

Now let me verify the non-patent literature citations and the sibling-application family.

Prior-Art Analysis — US 6,801,579 B1

Scope note / sourcing. I verified the citation record for US 6,801,579 against the Google Patents full-text record (https://patents.google.com/patent/US6801579/en), the USPTO Official Gazette entry (Indiana University virtual disk library, Issue 1287-1: https://webapp1.dlib.indiana.edu/virtual_disk_library/index.cgi/[5628977](/patent/5628977)/FID3294/OG/html/1287-1/us06801579-20041005.html), the Justia mirror, and the PatSnap "Eureka" record. I confirm the only patent citation on the face of 6801579 is US 6,724,842 B1, plus 14 non-patent citations and four cited co-pending Lucent applications. I did not retrieve full texts of the non-patent literature; anticipation conclusions below are therefore flagged as potential and rest on the abstract/scope descriptions I could verify, not on element-by-element comparison of every reference page. Everything here builds on the bibliographic and claim summaries already generated; where I extend or refine them I say so.

Date caveat (carried forward, not resolved). The task header says April 26, 2026; the operative system date is 2026-09-28; the patent fetch was stamped 2026-03-26. The prior sections already flagged this. It does not change the prior-art analysis, because § 102 provable dates here are all fixed patent/publication dates, and the patent is expired (adjusted expiration 2022-12-21).

---

1. Applicable statutory framework and critical dates

US 6,801,579 was filed 2000-08-22 and claims priority to provisional 60/188,137, filed 2000-03-09 (confirmed in the OG entry: "Claims priority of provisional application 60/188137, filed on Mar. 09, 2000"). It is therefore governed by pre-AIA 35 U.S.C. § 102 (pre-AIA §§ 102(a), (b), (e), and (f), and pre-AIA § 103(c)).

The candidate critical dates are:

This distinction is dispositive for the Hughes WCNC paper (Sept. 21–24, 1999): it falls after both critical dates, so it is not § 102(b) art. It can only be § 102(a) art (publication before the applicant's invention date, presumed to be the 2000-03-09 priority date). Everything published before 1999-03-09 is § 102(b) as well as § 102(a). Patent art is governed by § 102(e) (US patent granted on an application filed before the applicant's invention date).

Important structural observation that runs through this whole analysis: the on-face art is dominated by the applicants' own and their close Bell Labs colleagues' work. Two of the four named inventors (Hochwald, Sweldens) appear as authors/inventors on much of the cited material; Hassibi and Shokrollahi appear on the rest. This is a self-citation-heavy record, which matters both for § 102(a) timing arguments and for pre-AIA § 103(c) common-ownership disqualification.

---

2. The references, one by one

2.1 The sole patent citation

US 6,724,842 B1 — "Method for wireless differential communication using multiple transmitter antennas"

• Inventors: Bertrand Hochwald, Wim Sweldens (both co-inventors of 6801579). Assignee: Lucent Technologies Inc.
• Filed 1999-07-16 as Appl. No. 09/356,387; issued 2004-04-20; 27 claims. Verified at https://webapp1.dlib.indiana.edu/virtual_disk_library/index.cgi/5628977/FID3294/OG/html/1281-3/us06724842-20040420.html and https://patents.google.com/patent/[US6724842B1](/patent/US6724842B1)/en. EP family member EP 1 069 723 B1 (filed 2000-07-03; priority US 356387).
• Description (verified from the EP equivalent's specification): differential modulation for a multiple-transmitter antenna array in which each baseband signal is a transmission matrix (columns = antennas, rows = time segments) generated from a previous transmission matrix and an input data sequence. It expressly addresses the case "when the constellation forms a group" (EP1069723B1, Fig. 3), discloses diagonal codes ("every antenna transmits a PSK symbol in turn"), and includes a table of constellation searches for 2–5 transmit antennas. Confirmed quote: "…we immediately see why it is useful in practice to have ν form a group under matrix multiplication: if ν is a group then all the transmitted matrices S_τ also belong to ν." (This same statement appears verbatim in the 6801579 specification — the two share a common text block.)
• § 102 basis: pre-AIA § 102(e) — its 1999-07-16 U.S. filing date precedes the 2000-03-09 priority date. (Same-day nuance: because Sweldens is a common inventor, "by another" is satisfied only because the inventive entities differ — Hochwald+Sweldens vs. Hassibi+Hochwald+Shokrollahi+Sweldens. For obviousness purposes, pre-AIA § 103(c)(1) common ownership by Lucent is a potential disqualifier; for anticipation under § 102(e) it is not.)
• Claim mapping (potential): preamble of claims 1, 8, 9, 14, 21, 22 (selecting a unitary-matrix signal from a constellation and transmitting complex amplitudes from two or more antennas across successive time intervals), and the group-structured-constellation element of claim 1(I) and its receiver mirror claim 14(I). Its differential-encoding limitation is narrower than claim 1, which is not limited to differential transmission — so it is closer to a § 103 combination reference than a clean § 102 anticipation of the genus.

2.2 Non-patent citations (14)

---

3. Most relevant prior art — ranked, with claim mapping

Rank 1 — US 6,724,842 B1 (Hochwald & Sweldens, Lucent)

The only on-face patent reference, and the closest structural match to the claimed invention's transmission model. It discloses (a) multiple transmit antennas, (b) unitary space-time signal matrices with antennas as columns and time as rows, (c) differential encoding by matrix multiplication, and (d) explicitly the group-constellation case. It maps most directly against:

• Claim 1 / claim 14 — preamble (steps (a)–(c)) and the group-constellation element (I)/(III). Its Fig. 3 group treatment is the strongest single disclosure supporting the "constellation forms a group" limitation.
• Claim 8 / claim 21 — its diagonal codes disclose the A matrix (diagonal, root-of-unity powers) but not, as far as I can verify, the combined A^l B^k shift-generator structure.
• Claim 9 / claim 22 — its differential matrix-product encoding touches the "pairwise products" structure, but the cited art is directed to products of successive transmitted matrices, not to the two-representation product set S_{A,B} claimed.

Caveat in the applicant's favor: the reference claims differential communication; claim 1 is not so limited, so 6,724,842 is a stronger § 103 reference than a clean § 102 anticipation.

Rank 2 — Hughes, "Differential Spacetime Modulation," WCNC 1999 (§ 102(a))

This is the most legally consequential non-patent reference, and the reason is visible on the face of claim 1 itself. Hughes discloses "a general approach to differential modulation for multiple transmit antennas based on group codes… applicable to any number of transmit antennas and any signal constellation" with modulators designed for two transmit antennas (verified abstract, DOI 10.1109/WCNC.1999.797803, pp. 145–149). The 6801579 specification itself concedes that the quaternion groups D_{2n,1,−1} "were proposed for use in wireless communication in B. L. Hughes" — and claim 1 then carves out "a non-Abelian group, other than a quaternion group."

That carve-out is the classic signature of an applicant distinguishing the closest art. Consequence: Hughes's own quaternion-group disclosure cannot anticipate claims 1 or 14, because those claims expressly exclude it — but Hughes remains the closest art for the broad group-code genus and for the subset/multiplicative-closure and coset constructs in claim 1(II)–(IV). Its timing (§ 102(a) only, since Sept. 1999 post-dates both one-year critical dates) limits it, and a § 102(a) challenge requires the petitioner to establish the invention date.

Note also the later Hughes paper ("Further results on differential space-time modulation," https://ieeexplore.ieee.org/document/[877990](/patent/877990)/) showing every full-rank unitary group code is equivalent to a cyclic or dicyclic group code — that is the art that most closely frames claims 2–3 (the G_{m,r} and D_{m,r,l} classes), though the published version post-dates the priority date and is therefore not itself available as prior art.

Rank 3 — Hochwald et al., "Systematic Design of Unitary Space-Time Constellations," Bell Labs TM, Jan. 1999 (§ 102(b))

Published in the USTM lineage (the corresponding IEEE version is IEEE Trans. Info. Theory 46(6):1962–1973, Sept. 2000). This is the base unitary-space-time-constellation design art that the preamble of claims 1/8/9/14/21/22 presupposes. It is § 102(b) art on either critical date. Its significance is as a § 103 foundational reference, not a standalone anticipator of the group-structure limitations.

Rank 4 — Hochwald & Marzetta, "Unitary space-time modulation for multiple-antenna communications in Rayleigh flat fading," IEEE Trans. Info. Theory 46(2):543–564 (March 2000)

Verified: "Unitary space-time modulation uses constellations of T×M space-time signals… The columns of each Φ_l are orthonormal." This is the canonical USTM reference and the parent of the preamble of essentially every independent claim. Timing problem: "March 2000" publication is essentially simultaneous with the 2000-03-09 priority date, and the patent cites the Bell Labs TM predecessors instead (items 11–13 below). Treat it as § 102(a)/§ 103 background, not a clean § 102(b) reference.

Rank 5 — Tarokh et al., "A differential detection scheme for transmit diversity" (2000) — for claims 12 and 25 specifically

This is the art that the Hamiltonian claims (12 and 25) were drafted against. The 6801579 specification reproduces the Tarokh "orthogonal design":

OD(x,y) = (1/√2) [ x  −y* ]
                 [ y   x* ]

and states that the orthogonal construction requires |x| = |y| ("For orthogonal constructions, x and y must have equal absolute values"). Claim 12 then requires the opposite — a set including at least one matrix with ‖x| − |y‖ > 0 (claim 13: ≥ 0.1). Consequence: Tarokh's orthogonal design cannot anticipate claims 12/25, because it teaches the equal-modulus limitation the claims exclude. Tarokh 2000 is thus relevant as the distinguishing reference, not an anticipatory one — and its own "to be published" status (July 2000 publication) makes it marginal even as § 102(a) art.

Rank 6 — Tarokh et al., IEEE Trans. Info. Theory 44:744–765 (1998) (§ 102(b))

The space-time-code performance-criterion paper that supplied the rank/determinant/trace design criteria underlying the 6801579 "diversity product" metric. Highly relevant to the motivation behind claims 1–7 but discloses trellis/known-channel codes, not group-structured unitary constellations; not an anticipator.

Rank 7 — Hamkins et al. (1997); Conway & Sloane (1988); Devroye (1986)

All § 102(b) printed publications, all cited only for the Hamiltonian-constellation theory in the specification: dense spherical codes on the 3-sphere (Hamkins; Conway & Sloane) and the bucketing/nearest-point decoding technique (Devroye). Relevant as § 103 secondary references against claims 12/25, and as evidence that the sphere-packing goal was known — but none discloses the claimed |x| ≠ |y| matrix set.

Rank 8 — Hassibi et al., "Multiple antennas and representation theory," Bell Labs TM, Mar. 2000

This is the inventors' own representation-theory treatment (later published as Shokrollahi, Hassibi, Hochwald & Sweldens, "Representation theory for high-rate multiple-antenna code design," IEEE Trans. Info. Theory 47(6):2335–2367, Sept. 2001 — the "Shokrollahi et al., 2002" item in the Similar Documents list). Because it is substantively the same subject matter as claims 2–7 (the six fixed-point-free group classes), it is the single most conceptually relevant reference to the dependent claims. Its timing is the problem: March 2000 publication coincides with the 2000-03-09 priority date, and it is the applicants' own work, so it is not cleanly available under § 102(a) or (b). Flag it as the reason claims 2–7 may be narrow but not obviously novel — the applicants published the underlying mathematics essentially contemporaneously with the provisional.

Rank 9 — the cited co-pending sibling applications

• Ser. No. 09/134,297 (Aug. 14, 1998) → issued as US 6,327,310 B1, "Wireless transmission method for antenna arrays, having improved resistance to fading" (Hochwald & Marzetta; verified via EP 0981209 / Unified Patents record, application number 09134297). § 102(e) date 1998-08-14 — the earliest patent-art date in the record.
• Ser. No. 09/206,843 (Dec. 7, 1998) → US counterpart of EP 1 009 124 A3, "Wireless transmission method for antenna arrays using unitary space-time signals." § 102(e). (I did not confirm the U.S. patent number for this application — flagged as unverified rather than guessed.)
• Ser. No. 09/356,387 (Jul. 16, 1999) → US 6,724,842 B1 (item 2.1).
• Ser. No. 09/528,973 (Mar. 21, 2000) → issued as US 6,693,976 B1, "Method of wireless communication using structured unitary space-time signal constellations" (title match confirmed from the US 6,693,976 B1 PDF). Critically, this application was filed twelve days AFTER the 6801579 provisional and therefore is not § 102(e) prior art to 6801579. It is a co-pending sibling, not a reference. (The US 6,693,976 ← 09/528,973 mapping is inferred from the matching title; the application number on the PDF's front page was not captured in my source — flagged.)

---

4. Claim-by-claim § 102 summary

---

5. Bottom line and confidence

• The most relevant prior art to US 6,801,579 is (1) US 6,724,842 B1 — the sole patent citation, § 102(e), filed 1999-07-16 by co-inventors Hochwald and Sweldens — and (2) Hughes, "Differential Spacetime Modulation," WCNC 1999, § 102(a) only, which is the art the claim 1 and claim 14 "other than a quaternion group" carve-outs were evidently drafted to avoid.
• No on-face reference anticipates the genus claims cleanly, and none anticipates the dependent group-class claims 4–7 or the Hamiltonian claims 12/25. The strongest § 103 story would combine US 6,724,842 (group-structured differential unitary space-time signals) with Hughes (group codes) and the Hochwald USTM references (constellation design), but pre-AIA § 103(c)(1) common ownership by Lucent limits use of the Lucent-family references (US 6,724,842; App. 09/134,297; App. 09/206,843; App. 09/528,973) for obviousness.
• Confidence: high on the citation set and filing/issue dates (verified against two independent sources). Moderate-to-low on the element-level anticipation conclusions for the non-patent literature, because I did not read the full texts of Hughes WCNC 1999, Hamkins 1997, or the two Bell Labs Technical Memoranda. A definitive § 102 chart requires the reference texts.

Two flags I will not auto-correct or paper over:

1. Claim 1(II) recites "guaternion group" while claim 1(I) recites "quaternion group." This is a literal claim-scope irregularity (already noted in the prior claim summary). It affects which art can be said to read on which sub-element and is worth preserving verbatim in any chart.
2. The Hassibi Bell Labs TM (Mar. 2000) is by the 6801579 inventors themselves and coincides with the priority date. Any attempt to use it as prior art against claims 2–7 runs into both the § 102(a) timing question and the self-disclosure problem.

I'll ground this obviousness analysis in the actual prior-art references on the patent's face before building the combinations.

Good grounding. Let me verify two more points: the closest prior art's group-code content, and the orthogonal-design reference underpinning the Hamiltonian claims.

Obviousness Analysis — US 6,801,579 B1 under 35 U.S.C. § 103

Scope note. This builds on the prior summary, litigation, and PTAB sections (bibliographic data, claim breakdown, the "guaternion" typo, and the zero-challenge record) and does not repeat them. It anchors to the prior art on the face of the patent record per the "Patent Citations (1)" and "Non-Patent Citations (14)" tables at https://patents.google.com/patent/US6801579/en, supplemented by the reference texts I could reach. The current date for this task is April 26, 2026; note the operative system date is 2026-09-28 — the inconsistency flagged in the earlier litigation section persists and I do not resolve it here.

---

1. Bottom line

There is a colorable but not clean § 103 case against the broad genus claims (1/14), a weaker case against the set/group-product claims (8–11/21–24), only a moderate case against the two-antenna "Hamiltonian" claims (12/25), and a poor case against the six class-limiting dependent claims (2–7/15–20). The single most damaging fact for any challenger is structural, not evidentiary: the reference that discloses the actual heart of the invention — the classification of fixed-point-free groups and their irreducible unitary representations — is the inventors' own 2000 Bell Labs technical memorandum / ISIT paper, which is not "by another" and is within the grace period, and is therefore not prior art at all. The challenger is left with the framework art (Hughes; Hochwald/Sweldens; Hochwald/Marzetta) but must bridge the gap to the claimed group classes with general mathematical knowledge and KSR-style "finite number of predictable solutions" reasoning.

Practical caveat, stated once: the patent expired 2022-12-21 (adjusted), so this is a hypothetical/defensive exercise; see the prior PTAB section for why § 315(e)(2) estoppel is a non-issue.

---

2. Which statutory regime applies, and the critical dates

• Filed 2000-08-22, priority to provisional 60/188,137 filed 2000-03-09. This is pre-AIA § 103. The AIA first-inventor-to-file provisions do not apply.
• § 102(b) critical date: one year before the earliest effective U.S. filing date. If the claims are entitled to the 2000-03-09 provisional, the date is 1999-03-09; if not, 1999-08-22. Either way, a January 1999 publication is § 102(b) art.
• Grace period: documents published between the critical date and 2000-03-09 (e.g., Hughes, WCNC, Sept. 21–24, 1999) are not § 102(b) art, but are available under § 102(a) as printed publications by another predating the March 2000 invention date.
• § 103(c)(1) (pre-AIA) matters enormously here. Subject matter that qualifies as prior art only under § 102(e), (f), or (g) is disqualified for § 103 if commonly owned at the time the invention was made. Both the '579 and US 6,724,842 B1 were Lucent Technologies. Applied literally, the sole U.S. patent cited against the '579 is not available as § 103 art — it remains only § 102(e) anticipation art. Same result for the four co-pending Lucent applications cited in the NPL list (09/134,297; 09/206,843; 09/356,387; 09/528,973).

---

3. The prior-art universe as it actually stands

The asymmetry. The '579's specification is, in substance, the P12 disclosure. P12 cannot be used against it. Everything the challenger can actually deploy (P2, P4–P10) is the surrounding framework — the group-code idea and the unitary-constellation idea — not the group classification.

---

4. Person having ordinary skill in the art (PHOSITA)

For § 103 purposes I assume a PHOSITA as of March 2000 with (i) an M.S. or Ph.D. in electrical engineering / communications, or an M.S. in mathematics with communications exposure; (ii) 2–3 years of work in multi-antenna (MIMO) transmission, space-time coding and fading-channel modeling; and (iii) working familiarity with the rank/determinant criteria of P4 and the differential DPSK generalization of P2/P8, but not research-level finite-group representation theory. This is deliberately a conservative (challenger-unfriendly) construction: it makes the "would a POSITA have looked to the fixed-point-free group classification?" step harder.

Under a looser construction — PHOSITA includes someone with graduate algebra — the § 103 case against claims 1/14 strengthens materially.

---

5. Claim-construction points that move the obviousness analysis

1. Claim 1(I)/(II) excludes quaternion groups explicitly; (II) says "guaternion" (typo per the earlier summary, not auto-corrected). If "guaternion" is read as a typo for "quaternion," claim 1(II) excludes quaternion subsets. If it is not, claim 1(II)'s exclusion is inoperative and claim 1 becomes broader — a subset of a quaternion group with positive diversity product (Hughes' quaternion code, P2) would then satisfy (II). This is a § 112/§ 103 pincer: the same drafting defect either narrows nothing (indefinite) or broadens the claim into P2's disclosure.
2. Claim 1 is a disjunctive "at least one of (I)–(IV)" list. Invalidity therefore needs one reference or combination reaching one branch. A challenger need not prove (II), (III) and (IV).
3. Claim 8/21 "B" is described inconsistently — called a "diagonal matrix" though the depicted matrix is a companion/cyclic-shift matrix; claim 8 says "i = −1" while claim 21 says "i² = −1." For § 103 I construe B by its depicted structure (the shift matrix with a root-of-unity in the lower-left), which is what the specification's S_{m,s} construction requires.
4. Claim 12's "‖x|−|y‖ > 0" is the entire point of the Hamiltonian claims — the orthogonal-design constraint being removed.

---

6. Ground-by-ground analysis

6.1 Ground A — Claims 1 and 14 (the non-Abelian, non-quaternion group genus)

Proposed combination: P2 (Hughes WCNC '99) as the primary reference, in view of P6 (Hochwald et al., Jan. 1999) and P7 (Hochwald & Marzetta USTM), optionally with P4 (Tarokh 1998) and the classical fixed-point-free group classification.

Motivation to combine (challenger's story, KSR rationales):

1. Design need, stated in the references themselves. P2 explicitly derives "error bounds, and modulator design criteria, which we use to design optimal modulators." P6's stated purpose is to systematically design constellations with low inter-signal correlation. Both frame constellation selection as an optimization problem with a known objective (maximize the diversity/product distance). That is a textbook "the claimed invention is a predictable solution to a recognized problem."
2. Group structure was already the accepted answer. P2 teaches that group structure yields equal-energy codewords, congruent decision regions, and a simple differential rule — and table-lookup decoding (the '579's own stated practical advantage). A POSITA would therefore stay inside the group framework.
3. The excluded quaternion family was known to be sub-optimal in the regime that matters. P2's own characterization result — that for t = 2 and |G| = 2^p every full-rank unitary group code is equivalent to a cyclic or dicyclic code — was understood to be a restriction of the search, not an exhaustion of good codes. As constellation size grows, the quaternion/dicyclic codes degrade (the '579's Table 2 and Fig. 2 report the quaternion Q_6 losing by ~13 dB to SL₂(F₅) at R ≈ 3.45). A POSITA optimizing error rate at higher rate would look outside the cyclic/dicyclic pair.
4. Finite, enumerated solution space. The classification of finite fixed-point-free groups into six families (cyclic, quaternion/dicyclic, and the four additional families) was classical mathematics that long predated the filing (Burnside, Zassenhaus, Amitsur; survey literature). Combined with the known fact that a finite matrix group has full diversity iff it is fixed-point-free, a POSITA had a bounded list of candidates, each with a reasonable expectation of positive diversity product — i.e., a "finite number of identified, predictable solutions" (KSR).
5. P1/P6 teach searching explicitly. EP 1 069 723 A2 (the '842's EP counterpart) figures a table of "searches for constellations for 2, 3, 4, and 5 transmitter antenna arrays," and P6 describes the algebraic (coding-theoretic) route to generating constellations. Running such a search over non-Abelian non-quaternion fixed-point-free group representations is the natural next step.

Where the combination is weakest (and this is where the patent owner wins):

• Hughes' theorem arguably teaches away. For the commercially central case (t = 2, power-of-two constellation size), the prior art affirmatively stated that every full-rank unitary group code is cyclic or dicyclic. That is a teaching that, within that regime, non-quaternion non-Abelian groups are not available. The '579's counterexample (SL₂(F₅), 120 elements — neither a power of two nor dicyclic) lives outside Hughes' theorem's premises, but a challenger must show a POSITA would have realized the theorem's limits and gone hunting in the six-family classification.
• The bridge is mathematics, not references. None of P1–P11 teaches that "a group whose differences all have nonzero determinant must be fixed-point-free," nor enumerates the six families. On the conservative PHOSITA of § 4, that is a genuine, non-trivial insight — and it is exactly the '579's stated contribution.
• § 103(c) removes P1. The most natural "search" teaching (P1/EP 1 069 723) is commonly owned Lucent subject matter qualifying only under § 102(e) — disqualified for § 103.
• So Ground A reduces to P2 + P6 + [general knowledge of fixed-point-free group classification] — a two-reference-plus-knowledge combination whose third leg is the contested step.

Assessment: ~45–55% probability of a challenger prevailing on claims 1/14 in a contested proceeding, dependent almost entirely on whether the tribunal treats the fixed-point-free classification as "knowledge of the art" (KSR "predictable solutions") or as the inventive step.

---

6.2 Ground B — Claims 8 and 21 (the A^l B^k / S_{m,s} construction)

Proposed combination: P2 (Hughes) in view of P6 (Hochwald et al.), optionally with P1.

Why this is a genuinely strong ground on the structural elements:

• A is squarely disclosed. Hughes' cyclic group code takes G = diag(ω_M, ω_M^k) with ω_M = exp(2πj/M) — a diagonal matrix each of whose entries is an integer power of an M-th root of unity. That is claim 8's "A," element (II), almost verbatim.
• B is squarely disclosed, up to conjugation. Hughes' dicyclic generator is [[0,−1],[1,0]] — an n×n companion/shift matrix with a root-of-unity (here ±1) in the lower-left, i.e. claim 8's "B," element (III), with s = 1 or 2. Together, A and B generate the group via BAB^{-1} = A^r — the same presentation (σ^m = 1, τ^n = σ^t, στ = σ^r) that the '579 uses for G_{m,r}.
• Generalizing 2×2 to n×n, and letting s be free of n, is routine engineering. Claim 8(IV) merely frees s from being tied to n (that is precisely the G_{m,r} → S_{m,s} generalization). "More design freedom so as to search for a higher diversity product" is an articulated rationale, and Hughes himself says the group-code approach "can be applied to any number of transmit antennas and any signal constellation."
• Claim 21 is the receiver mirror and is met by P2's correlation receiver (arg max Re Tr{G Y_k† Y_{k−1}}) when the constellation is the A/B-generated set.

Assessment: ~60–65%. The element-by-element overlap with Hughes' two explicit generators is close enough that the only real dispute is whether the claim's added generality (n arbitrary, s free / p = min(s,n)) carries patentable weight. Under KSR, an obvious generalization of a disclosed species is not.

---

6.3 Ground C — Claims 9, 22 (and 10/11, 23/24) — pairwise products of two group representations

Proposed combination: P2/P6 as the group-code framework, plus general group theory (direct products of unitary representations; Kronecker products).

• The framework is suggested: P6's construction generates constellations by repeated multiplication by a unitary matrix Θ (including Θ a root of the identity), and P2 generates by group multiplication. A set of pairwise products of two representation sets is a natural composite.
• But there is a genuine specificity problem. The '579's own dependent claims 11/24 limit this to the counter-intuitive case where G_A = G_B as equivalent reducible representations of an odd-order group — and the specification states that this yields positive diversity product only under those conditions. I found no teaching in P1–P11 of this T A_j T* construction or of the odd-order/reducibility precondition.
• Assessment: independent claims 9/22 ~35–45% (mere aggregation of known group-code sets, arguably obvious); dependent claims 11/24 materially stronger for the patent owner (~15–25%), because the claimed sub-case is a non-obvious mathematical refinement.

---

6.4 Ground D — Claims 12 and 25 (Hamiltonian constellations, ‖x|−|y‖ > 0)

Proposed combination: P5 (Tarokh & Jafarkhani orthogonal design OD(x,y) = (1/√2)[[x, −y*],[y, x*]], x, y unit-modulus and typically Q-th roots of unity) in view of P9 (Conway & Sloane sphere packings) and P10 (Hamkins & Zeger, asymptotically dense spherical codes), and P6 (minimum-distance/unitary design).

The case for obviousness is reasonably clean here, and it is the one ground where the patent's own reasoning supplies the bridge:

• Claim 12's matrix form is identical to P5's orthogonal design. The '579 specification says so: "According to that rule, which we refer to as an orthogonal design, each signal matrix takes the form OD(x,y) …"
• P5's constraint is |x| = |y| (unit modulus). Claim 12 requires |x|² + |y|² = 1 with at least one matrix having ‖x| − |y‖ > 0 — i.e., the points are allowed off the Clifford torus {|x| = |y| = 1/√2} and onto the rest of the unit 3-sphere.
• P9 and P10 are expressly about packing that very sphere. P10 ("asymptotically dense spherical codes") and P9 (sphere packings in S³) both teach the design principle the '579 invokes: pick L points on the unit 3-sphere to maximize minimum Euclidean distance. Both are on the face of the patent as cited NPL (so presumptively the examiner saw them).
• The mathematical result is straightforward arithmetic: the OD family has ζ_OD = sin(π/Q)/√2 = O(L^{−1/2}), whereas a good spherical packing of S³ gives O(L^{−1/3}). A POSITA with P9/P10 in hand would naturally ask whether the torus restriction was costing distance. That is a classic "recognized the limitation, applied a known packing theory to remove it" fact pattern.
• Claim 13/26's "at least 0.1" is a quantitative floor that would be met by essentially any genuinely non-toroidal sample point; it adds little.

Weakness: a challenger must show motivation to abandon orthogonality, since the orthogonal design's whole selling point was structural simplicity and (for the coherent Alamouti case) orthogonality/decorrelation. On the conservative PHOSITA of § 4, expecting him to import the sphere-packing literature into the differential 2×2 design is a real stretch — though the '579 itself argues the payoff is asymptotic, which cuts toward "predictable from the packing literature."

Assessment: ~50–60%.

---

6.5 Ground E — Dependent claims 2–7 and 15–20 (the six group classes)

Assessment: poor for the challenger (~10–20%). These claims narrow to specific families — G_{m,r} (order mn, n antennas), D_{m,r,l} (order 2mn, 2n antennas), E_{m,r} (8mn), F_{m,r,l} (16mn), J_{m,r} (120mn), K_{m,r,l} (240mn) — with prescribed group-order-to-antenna-count relationships and, by claim, n the multiplicative order of r mod m.

• None of E, F, J (≅ SL₂(F₅) × G_{m,r}), or K appears anywhere in P1–P12 or the earlier literature. P2 discloses only cyclic and dicyclic families. P1/P6 disclose no group taxonomy.
• The classification of fixed-point-free groups into these six types, and their irreducible fixed-point-free unitary representations (the A, B, P, Q, R matrices of the specification), is the P12 disclosure — not prior art (same inventive entity; within grace period). See § 6.6.
• Notably, the interesting SL₂(F₅) example (120 elements, ζ = 0.3090, the best-performing constellation in the patent) is J_{1,1} — a member of a family the prior art did not name.

The washout here is the counterpart of the "blank slate / immune" point made in the earlier litigation/PTAB sections: a challenger who wins the genus may still fail wholesale on the species.

---

6.6 The load-bearing point: P12 is not available

I return to this because it determines the whole analysis. The only pre-filing document that discloses (i) the fixed-point-free property as the necessary condition for full diversity, (ii) the six-family classification, and (iii) the explicit unitary matrix representations of each family, is Hassibi, Hochwald, Shokrollahi & Sweldens, "Multiple antennas and representation theory" — the same four inventive entities as the '579. Under pre-AIA § 102(a) it is not "by another"; and it published roughly contemporaneously with, and in the case of the Bell Labs TM possibly before, the March 9, 2000 provisional, so § 102(b)'s one-year requirement is also unmet. It cannot be used for either anticipation or obviousness, and it cannot be used as the "reason" that supplies the missing bridge in Grounds A–C.

I flagged the corollary risk in the earlier section: if the provisional (or the Bell Labs TM's actual public-availability date) ever undermined the priority claim such that the invention date slipped past the March 2000 memo, the entire analysis inverts and the '579 falls to its own authors' paper. That is the single highest-value factual question for either side, and the one I cannot resolve from the record.

---

7. Conspectus of grounds

---

8. Countervailing considerations the challenger must answer

1. Objective indicia of non-obviousness. The specification reports SL₂(F₅) beating the best orthogonal design by ≈2.5 dB, the best Abelian group by ≈6.5 dB, and the quaternion group by ≈13 dB, plus a real over-the-air demonstration in a hallway (Fig. 6). Unexpected results and failure of others are the natural rebuttal. The weakness of this evidence: it is entirely in-spec, generated by the inventors, and there is no third-party licensing or industry-adoption evidence in the record (consistent with the earlier finding of zero assertion activity).
2. Teaching away (Hughes' theorem). P2 affirmatively characterized the t = 2 power-of-two universe as cyclic-or-dicyclic only. Combined with the fact that the quaternion code is optimal in the small-constellation 2-antenna regime (ζ = 0.7071), the references point toward the excluded family, not away.
3. § 103(c) common ownership knocks out P1 and the four co-pending Lucent applications, forcing the challenger onto third-party art (P2, P4, P5, P9, P10) plus mathematical knowledge.
4. Claim-drafting defenses. Claim 1's disjunctive structure and the "guaternion" typo interact: a challenger attacking (II) faces indefiniteness (not invalidity) and a patent owner who can disavow the typo in litigation and keep (I).

---

9. Verification gaps I could not close

• I have the abstracts, theorem statements, and figure captions of P2 and the Hughes/Hochwald group-code papers, not verified full text of the Sept. 1999 WCNC paper's § III.A. Whether Hughes expressly discloses any non-quaternion non-Abelian group with positive diversity product is the single fact that would flip Ground A from contestable to strong. I could not confirm it either way.
• I did not obtain the full text of P1 (US 6,724,842 B1) or EP 1 069 723 A2 to confirm precisely what the Fig. 4 "search" table discloses (candidate groups vs. only diagonal/cyclic).
• The public-availability date of the March 2000 Bell Labs technical memorandum (P12) is imprecise in the record; this drives the priority question in § 6.6.
• My search sequence was truncated before I could retrieve P5 and P9/P10 in full.

Suggested confirmatory steps: (1) pull the Hughes WCNC '99 paper § III and any pre-March-2000 manuscript version to see whether non-quaternion non-Abelian groups are named; (2) pull P1/EP 1 069 723 A2 Fig. 4 disclosure; (3) establish the exact date the Bell Labs TM "Multiple antennas and representation theory" became publicly accessible; (4) confirm the provisional's disclosure scope against the issued claims (this is the priority question on which the whole § 103 posture turns).

---

10. Bottom line

Under pre-AIA § 103, the strongest realistic attack on US 6,801,579 is Hughes (WCNC '99) + Hochwald et al. (Jan. 1999 Bell Labs TM) + Hochwald/Marzetta USTM, argued as the predictable exploitation of the (classical) fixed-point-free group classification to improve the diversity product of a differential multi-antenna group code — with KSR rationales of a stated design need, an already-adopted group framework, and a finite, enumerable solution set. That ground has a fair chance against claims 8/21 and a coin-flip chance against claims 1/14, and it is met by the "Hamiltonian" case (12/25) from the orthogonal-design art plus the spherical-packing literature. It fails against the six class-limiting dependents (2–7/15–20) and probably against the equivalent-reducible-representation dependents (11/24), because those species trace to the inventors' own 2000 representation-theory work — which, being the same inventive entity and within the grace period, is not prior art at all.

That structural feature — the invention's core disclosure was authored by the inventors and is therefore unusable, while the surrounding framework art is usable but incomplete — is what makes this patent, in the end, look more robust on obviousness than the crowded 4G/5G lineage and the total absence of any challenge might superficially suggest. It remains, however, untested: the earlier sections correctly establish there is no IPR, no reexam, and no litigation in which any of this has ever been litigated, so every figure above is a prediction, not a finding.