US 10644380

I'll run targeted searches against the USPTO and CAFC-facing sources for this exact number, plus pull the abstract and claim set I don't have verbatim.

Let me pull the abstract and claim set verbatim, and separately check CAFC 2026 dockets for Fractus.

Let me pull the abstract and the granted claim text directly, plus a final CAFC check.

Patent Summary — US 10,644,380 B2

Sources and scope note

I ran targeted searches for the literal string 10644380 against USPTO-hosted records (PTACTS petition documents, file receipts, ODP/Google Patents bibliographic block) and against CAFC-facing sources. I did not obtain an authenticated PatentCenter/PTAB E2E lookup or a complete CourtListener/PACER CAFC docket sweep — that limitation is carried forward from the earlier sections and I flag it again below.

Date conflicts I will not silently reconcile: this task's header states April 26, 2026; the top-level session header states 2026-09-28; the earlier Litigation section operated at 2026-09-27; and one source I retrieved is a petition filed April 20, 2026. These cannot all be simultaneously true. I report dates as each source literally gave them.

---

1. Bibliographic data

Naming inconsistencies to note (not auto-corrected):

• Inventor 3 appears as "Jordi ILARO" on the granted front page / Google Patents inventor field, and as "ILARIO, JORDI" in the recorded re-assignment text. Same person, two spellings.
• A PTAB document (the '103 patent's petition, IPR2026-00338) renders the family title in the singular — "Multiple-Body-Configuration Multimedia and Smartphone Multifunction Wireless Device." The '380's own title uses the plural. Do not treat these as the same literal string.
• One source (Unified Patents portal) lists the '380 priority date as 2006-07-17 rather than 2006-07-18; all other sources say 2006-07-18. I record the discrepancy.

---

2. Abstract

Uncertainty: I could not retrieve the verbatim "(57) Abstract" text of the '380, and the full patent text supplied to me is truncated before any Abstract section (it ends mid-sentence in the optimization discussion). I will therefore not reproduce an abstract as if it were the granted text.

What I can state with confidence is the substance the abstract is drawn from, taken from the patent's own Summary/"Definitions" section:

The invention relates to a multifunction wireless device (MFWD) — more particularly a handheld one — that combines mobile data and voice services with at least one of multimedia-terminal (MMT) and personal-computer/smartphone (SMRT) capabilities. The device includes an antenna system whose contour exhibits a defined level of geometric complexity, parameterized by complexity factors F21 and F32, so that the antenna operates efficiently across multiple frequency bands in at least three regions of the electromagnetic spectrum while fitting within the device's available antenna volume.

That is a faithful paraphrase of the specification, not a quotation of the abstract. Anyone needing the exact 150-word abstract should pull the PDF front page (patentimages.storage.googleapis.com/b3/2f/b1/a25621b753024b/US10644380.pdf).

---

3. Plain-language overview of the independent claims

The '380 presents three independent claim families (the specification's Summary is drafted in claim language and mirrors them). All three share the same antenna limitation: an antenna system whose contour complexity is bounded by F21 between 1.05 and 1.80 and F32 between 1.10 and 1.90. The families differ only in the device features recited.

Claim family A — the multi-body device. A multifunction wireless device with an upper body and a lower body adapted to move relative to one another in at least one of a clamshell, slide, or twist manner, plus an antenna system disposed in at least one of the bodies, the antenna having the F21/F32 complexity bounds.

Claim family B — the smartphone device. A multifunction wireless device with a microprocessor and operating system capable of running word-processing, spreadsheet, and slide (presentation) software applications, and at least one memory coupled to the microprocessor with a total capacity of at least 1 GB — plus the same complexity-bounded antenna system.

Claim family C — the multimedia-terminal device. A multifunction wireless device with a receiver of analog and/or digital sound signals, an image recording system that includes at least one of a ≥2 Megapixel image sensor, a flash, an optical zoom, and a digital zoom, and data storage of at least 1 GB — plus the same complexity-bounded antenna system.

Dependent claims narrow the complexity windows (e.g., F21 > 1.39 with F32 > 1.50; F21 > 1.43 with F32 > 1.63) and recite device thicknesses in the 5–9 mm range.

Literal-reading caveat. In the patent text supplied to me, claim family A's limitation reads: "…defined by complexity factors F21 having a value of at least 1.05 and not greater than 1.80 and having a value of at least 1.10 and not greater than 1.90." As written, that sentence names only F21 yet recites two numeric ranges. Families B and C state the limitation cleanly as "complexity factor F21 … at least 1.05 and not greater than 1.80 and complexity factor F32 … at least 1.10 and not greater than 1.90." I read the family-A sentence as a transcription/OCR artifact where the second range belongs to F32 — but I have not verified this against the granted claim text, and I am flagging it rather than correcting it. Confirm from the printed patent.

Related family distinction worth keeping straight: the child patent US 11,031,677 ('677) recites F21 ≥ 1.20 and F32 < 1.75, i.e., a narrower window than the '380's broader [1.05–1.80] / [1.10–1.90]. The '380's windows are the broader, and therefore the more difficult, limitation to design around.

---

4. USPTO and CAFC docket check for 10644380

No AIA trial proceeding (IPR/PGR/CBM) naming US 10,644,380 was found. Consistent with the earlier PTAB section: the '380 appears in PTAB papers only inside the "Continuing Data" chain (e.g., "16832820 is a CON of 15856626 12/28/2017 PAT 10644380"), never as a challenged patent.

No CAFC docket naming 10,644,380 was found. My CAFC-directed query ("Federal Circuit 2026 docket Fractus S.A. appeal") returned no appeal involving the '380 — which is structurally expected, since a CAFC appeal of this patent would require an adverse final written decision or district judgment on it, and I located neither. I could not complete an exhaustive CourtListener/PACER RECAP query within my available search steps, so treat this as a targeted-search negative, not a certified null.

New family-adjacent item found in this pass (updating the earlier PTAB section):

• IPR2026-00338, Johnson Controls Inc. v. Fractus, S.A., filed April 20, 2026, challenging US 8,738,103 — the '380's direct ancestor — as to claims 1–11, 13–14, and 16–20. Grounds: obviousness over Boireau (US 2007/0238483), CN456 (CN1728456A), and Ciais-Quadband (April 2004 journal article), with Cho (2005) added for dependent claims 5 and 11. Petitioner measured the Ciais-Quadband antenna at F21 = 1.31, F32 = 1.57. This is a new petitioner (Johnson Controls) beyond Geotab/ResMed/Verizon/Vivint. It again names a sibling, not the '380.

Also confirmed in this pass: the PGR2025-00056 petition by Geotab is directed to US 12,095,149 ('149), and the Fractus Patent Owner brief references Fractus, S.A. v. Geotab Inc., No. 2:24-cv-01008 and Fractus, S.A. v. Verizon Connect Inc., No. 2:24-cv-01009 (E.D. Tex.), with a Markman hearing March 11, 2026 and trial September 14, 2026. No '380 in those.

Bottom line for this search: for the literal number 10644380, the CAFC 2026 docket is empty and the PTAB record is empty. Activity exists only on its ancestors ('103, '458), its child ('677), and its descendants ('149, '200).

---

5. Uncertainty register

1. Verbatim abstract — NOT obtained. §2 is a specification-based paraphrase, clearly labeled.
2. Verbatim granted claim text — NOT obtained. §3 is drawn from the patent's own Summary plus PTAB exhibits; the family-A F21/F32 duplication issue and the exact count of independent claims should be verified against the printed patent. A file receipt in the record shows "TOT CLAIMS 20 / IND CLAIMS 3," but that receipt is for application 17/246,192 (the '200 child), not for 15/856,626 — so it is corroborative, not dispositive, as to the '380.
3. Expiration date — 2026-12-21 is a computed/anticipated date only. It is consistent with a terminal disclaimer tying the '380 to the 20-year term of the 2006-12-21-filed ancestor (US 8,738,103), but I have not verified the terminal disclaimer in the 15/856,626 file wrapper. If no terminal disclaimer exists, the '380 could instead run from its own 2017-12-28 filing — a materially different exposure analysis.
4. Source-date conflict (April 26, 2026 vs. Sept. 26–28, 2026) — flagged, unresolved.
5. Priority micro-discrepancy (2006-07-17 vs. 2006-07-18) — flagged.
6. CAFC null is a targeted-search result, not an authenticated docket check.

Authoritative next steps: USPTO PatentCenter for application 15/856,626 (full claim set, Abstract, terminal disclaimer, maintenance-fee status) and PTAB E2E for patent number 10,644,380; CourtListener RECAP / Docket Alarm filtered to "Fractus" as appellant for CAFC confirmation.

I'll search for litigation involving US patent 10644380 across patent litigation databases.

Let me search more targeted litigation databases for assertions of this specific patent.

Litigation Report — U.S. Patent No. 10,644,380

Bottom line

I could not identify any litigation (district court, ITC, or CAFC appeal) in which U.S. Patent No. 10,644,380 ("the '380 patent") is itself asserted as a patent-in-suit. My searches returned Fractus's enforcement campaigns, and the '380 patent appears in the resulting records only in the priority/continuity chain of asserted sibling patents (e.g., in PTAB petitions identifying the '677 and '200 patents as continuations of the '380), never as an asserted claim set.

I want to be explicit about the limits of this conclusion:

• The '380 patent is a Fractus, S.A. patent (inventors Carles Puente Baliarda, Josep Mumbru, Jordi Ilaro), issued May 5, 2020 from application 15/856,626 (filed Dec. 28, 2017). It is a continuation of the family that also produced U.S. 8,738,103; 11,031,677; 11,349,200; and 12,095,149 — several of which have been litigated.
• The Google Patents page for US10644380 carries a family-level "Family has litigation / First worldwide family litigation filed" flag (Darts-IP link). That flag reflects worldwide litigation involving the family, and is consistent with the siblings being litigated — it is not evidence that the '380 itself was asserted. I treat it as such.
• I was unable to complete an exhaustive docket-by-docket check (e.g., full CourtListener/PACER/Unified Patents queries specifically for the string "10,644,380") within the available search steps. Some lower-volume or recently filed matters could have been missed.

Related Fractus enforcement matters — none of which assert the '380

For context, these are the Fractus cases that surfaced. In every one, the patent-in-suit list I could verify does NOT include the '380.

Note: The Samsung appeal and the associated reexaminations (Fractus v. Samsung, E.D. Tex.) involved the older "multilevel" patents, not the '380.

Adjacent PTAB/reexam activity (family members, not the '380)

The '380 patent's siblings and continuations have been challenged at the USPTO, which may be the source of confusion:

• Geotab filed IPRs on U.S. 11,031,677 and PGR2025-00056 against another Fractus patent (the '677 is a continuation of the '380).
• Unified Patents obtained institution of ex parte reexaminations on 11,349,200 and 8,593,349; a later reexamination request on a Fractus patent was discretionarily denied (Law360, May 26, 2026).
• In the PTAB papers, the '380 patent appears only in the "Continuing Data" chain (e.g., "15856626 is a CON of 14738090 … PAT 10644380"), confirming it as an ancestor — not as the challenged or asserted patent.

Recommendation for confirmation

If you need a definitive negative:

1. Run a plaintiff-side docket search (CourtListener RECAP / Docket Alarm / Lex Machina) for "Fractus" as plaintiff and filter patents-in-suit for 10,644,380.
2. Check Google Patents' litigation tab and the Darts-IP link on the US10644380 page for the worldwide family listing.
3. Check Unified Patents' litigation portal for Fractus (they track Fractus's assertions closely) to see the full asserted-patent list across campaigns.

If litigation asserting the '380 patent exists, I did not find it, and I am not asserting with certainty that none exists — only that the verifiable Fractus complaints I located assert other family members, not this patent.

I'll start by verifying the structured proceedings list against live web sources, since the ODP block says no AIA trials are on file.

Let me drill into the specific patent and the Fractus IPR cluster.

The '380 itself doesn't appear yet — the challenges are hitting its continuation siblings. Let me verify directly.

AIA Trial Proceedings — US 10,644,380 B2

"Multiple-body-configuration multimedia and smartphone multifunction wireless devices"
Patent Owner: Fractus, S.A. | Priority: 2006-07-18 | Filed: 2017-12-28 | Granted: 2020-05-05
Analysis date: 2026-09-27

---

Proceedings overview

Total AIA trial proceedings on file against US 10,644,380: zero. The canonical USPTO ODP "PTAB proceedings on file" block for this patent is empty — no IPR, PGR, or CBM has ever been filed or instituted against the '380 patent — and my targeted web verification (USPTO PTACTS petition documents, Docket Alarm PTAB dockets, RPX Empower, Unified Patents' portal, and the Fractus-focused PTAB briefing that surfaced) produced no petition, institution decision, or FWD naming US 10,644,380. Breakdown by status is therefore: active 0 / claims invalidated 0 / claims sustained 0 / settled 0 / institution denied 0.

The bottom-line defensive posture is unusual and, frankly, better than a "hardened" patent would be: the '380 is wholly untested at the PTAB, but it is also nearly dead. Google Patents computes an anticipated expiration of 2026-12-21 — roughly 85 days from today. That date is consistent with a terminal disclaimer tying the '380 to the 20-year term of its ancestor U.S. 8,738,103 (filed 2006-12-21, from which the '380 claims priority through the chain '103 → 9,099,773 → 9,899,727 → '380). If that mapping holds, no IPR is worth the filing fee: the trial would not reach an FWD until well after the patent's death, and the enforceable damages window closes in December 2026. There is no "claims 1-5 have been canceled" story here, and equally no "survived two IPRs" story — this patent has never been tested.

---

Proceedings on US 10,644,380

None. I am explicitly declining to populate the per-proceeding template, because doing so would require fabricating a proceeding number, and I found no candidate. For the record, the reasons the '380 has never drawn an AIA filing are structural, not accidental:

• No PGR was possible. The '380 claims 2006-07-18 priority (U.S. provisional 60/831,544 and EP 06117352.2) through a long continuation chain. PGR under 35 U.S.C. §§ 321–329 reaches only patents with an effective filing date on or after 2013-03-16. The '380 is outside the PGR-eligible class entirely.
• The CBM window closed without a filing. Covered Business Method review under AIA § 18 was available, if at all, only between the '380's issuance (2020-05-05) and the 2020-09-16 sunset for new CBM petitions. No CBM was filed in that ~4.5-month window. (Whether the '380 — an antenna-structure patent — could even have satisfied the "financial product or service" / "technological invention" gate is doubtful; the point is moot.)
• IPR was the only live vehicle, and it was aimed at the siblings instead. As of 2026-09-27 I found no IPR petition against the '380.

If you need belt-and-suspenders confirmation, check the "Post Grant" / "PTAB" tab in USPTO PatentCenter for application 15/856,626 and query PTAB E2E for the patent number directly. I could not perform that authenticated lookup from here, so treat my "zero" as verified-by-targeted-public-search rather than verified-by-Office-record-api-alone.

---

Family-adjacent proceedings (NOT against US 10,644,380)

These are the reasons the '380 never needed to be challenged: petitioners went after its ancestors, descendants, and siblings, all in the same Fractus continuation family. They are set out here so you understand the litigation posture — but none of these has any estoppel or claim-level effect on the '380.

Two non-AIA Office proceedings in the same family are worth knowing about because they define the current risk profile for any challenger:

• Ex parte reexamination 90/019,275 (Unified Patents, on US 11,349,200) — request granted 2024-01-10 on substantial-new-question grounds. This is the only successful Office-side validity challenge I found in the family, and it is [not an AIA trial] and does not touch the '380.
• Ex parte reexamination 90/015,984 (Geotab, on US 8,810,458) — denied 2026-05-22 under 35 U.S.C. § 325(d), on the theory that it re-presented "the same or substantially the same prior art or arguments" from the discretionarily denied IPR2025-00928/00929.

Caveat on conflicting sourcing: My sources disagree on which of IPR2025-00928 and IPR2025-00929 maps to which Fractus patent. One practitioner analysis attributes the '458 patent to IPR2025-00928, while RPX Empower titles IPR2025-00929 as "IPR of '458"; a CourtListener filing lists both numbers jointly as "Geotab Inc. v. Fractus, S.A., IPR2025-00928, -00929 (Sep. 12, 2025)." Both were denied on 2025-09-12. I have not resolved the mismatch and will not guess.

---

Strategic summary

Claim status on the '380: everything is UNTESTED. Not one claim of US 10,644,380 has been canceled, confirmed, or even challenged in an AIA trial. Claims 1–20 (the '380's claim set) stand exactly as they issued on 2020-05-05, unadjudicated. Any defendant receiving a demand letter citing the '380 cannot point to an FWD canceling anything — the leverage available against this patent has to come from somewhere else: the § 102/103 record, the priority/terminal-disclaimer chain, and the clock.

Estoppel landscape: clean for everyone. Because no proceeding against the '380 has ever reached an FWD, no petitioner and no privy is subject to § 315(e)(2) estoppel as to this patent. The full universe of prior art is available to a defendant — nothing was "raised or reasonably could have been raised" before the Board on the '380, because nothing was. Note the asymmetry: the Geotab petitions against the '677 and '200 (IPR2025-01026, IPR2025-01027) may generate estoppel on those patents once/if they end in FWD, but that estoppel attaches to the patents challenged, not to the '380. Conversely, a defendant who files its own IPR on the '380 today would face a § 315(e)(2) bar in the parallel district case only to the extent a trial is actually instituted — and post-expiration, institution is a poor bet.

Pattern signals. Three patterns are worth flagging. (1) Geotab ran the mature-challenger playbook and lost on discretion twice, then lost again on reexam. It filed a contemporaneous wave (IPR2025-00928/-00929 on the '458, IPR2025-01026 on the '677, IPR2025-01027 on the '200, plus PGR2025-00056); the '458 pair was knocked out on the Acting Director's "settled expectations" reasoning (patent in force >11 years; petitioner had notice for years), and when Geotab tried the same art via ex parte reexam in February 2026, the CRU used § 325(d) to shut the door on 2026-05-22. That is a strong warning that copying a previously-filed Fractus challenge verbatim will be denied, even where the merits were never reached. (2) Fractus litigates rather than defends at the Board — it won IPR2026-00196 and IPR2024-00087/-00088 by settlement rather than by FWD, and its EDTX campaign (No. 2:24-cv-01009, consolidated with No. 2:24-cv-01008, Judge Gilstrap) resolved against both Verizon (settled March 2026) and Geotab (settlement announced 2026-07-27, terms confidential). (3) No defensive aggregator has touched the '380. Unified Patents is in the chain only via its ex parte reexam on the '200 (90/019,275) — again, not this patent. There is no Unified IPR on the '380 to free-ride on.

The single most important fact: the '380 dies on 2026-12-21. Google Patents' computed anticipated expiration is 2026-12-21, and the litigation history independently corroborates it — the Examiner rejected the related family members for obviousness-type double patenting over the '8,738,103 parent, and Fractus cured that with a terminal disclaimer, which by operation of law ties the disclaimed patent's term to the reference patent's term. The '103 was filed 2006-12-21. Twenty years later is exactly 2026-12-21. Verify this on PatentCenter (check the terminal disclaimer in the 15/856,626 file wrapper and the maintenance-fee status), but if it holds, the '380 has ~85 days of enforceable life and no independent term of its own.

---

Recommended next steps

1. Do not file an IPR on the '380 — and do not quote a canceled claim, because there isn't one. The honest defensive sentence for a client is: "US 10,644,380 has never been tested at the PTAB, so there is no FWD to hang your hat on; but it expires 2026-12-21, which caps exposure to pre-expiration sales and forecloses injunctive relief." Confirm the expiry first.
2. Verify the expiry before you do anything else. Pull the file wrapper for application 15/856,626 in PatentCenter and confirm (a) the terminal disclaimer over the '103, (b) payment of the 3.5-year and 7.5-year maintenance fees, and (c) any PTA. If the terminal disclaimer is absent and the '380 runs on its own 20-year term from 2017-12-28, your calculus changes materially and the patent may live to 2037-12-28 — in which case an IPR becomes a real option. I could not verify this from open sources and am flagging it as the decisive open question.
3. Identify which patent the demand actually cites. Fractus's active assertions run through the '677, '200, and '149 — the descendants of the '380 — in Fractus, S.A. v. Geotab Inc., No. 2:24-cv-01008 (E.D. Tex.) and Fractus, S.A. v. Verizon Connect Inc., No. 2:24-cv-01009 (E.D. Tex.), consolidated. If the letter cites a sibling rather than the '380, the defenses differ: those siblings are PTAB- and reexam-tested, and Geotab's IPR2025-01026/-01027 outcomes (which I could not confirm) will tell you whether the asserted claims survived.
4. If you do end up facing an actual '380 suit, calendar the § 315(b) bar immediately. The one-year IPR clock runs from service of a complaint alleging infringement of this patent. Service on other Fractus patents does not start it. But weigh any petition against the 2026-12-21 date — an FWD would land roughly 18 months after filing, i.e., long after the patent expires, leaving only cancellation-of-expired-claims value.
5. If a ground you are considering was already run by Geotab against a sibling, do not recycle it. The 2026-05-22 § 325(d) denial in EPR 90/015,984 shows the Office will refuse a challenge that merely re-presents prior art from a discretionarily denied petition — even one never examined on the merits. Either bring genuinely new art or bring it in district court, where § 325(d) does not apply.
6. Monitor for a late-breaking petition. Nothing forces a challenger to act before expiration, but if one is filed in the next ~85 days, institution and FWD dates will be governed by the statutory 1-year trial deadline from institution (35 U.S.C. § 316(a)(11)) — a timeline that necessarily runs past the patent's expiry.

Sources for the citations above: Google Patents record for the '380 and the ODP proceedings block (authoritative for count/status); USPTO PTACTS petition filings at ptacts.uspto.gov; Docket Alarm PTAB dockets (IPR2018-01461, IPR2026-00196); RPX Empower, IPR2025-00929; scalejuris analysis of the 2026-05-22 § 325(d) reexam denial; Unified Patents portal, ex parte reexam 90/019,275; Fractus/Geotab settlement announcement, 2026-07-27. Federal Circuit docket materials are available via CourtListener — no CAFC appeal involves the '380, because no FWD on the '380 exists to appeal.

I'll research the assignment record for US 10,644,380 across the available sources.

Let me dig into the specific assignment records and reel/frame numbers.

Assignment Reconstruction — US 10,644,380 B2

Method note / honest limitations. I was able to reach the Google Patents legal-events record (fetched 2026‑09‑26) and the patent's own front-page data, plus Fractus's litigation and portfolio history. I could not open the USPTO Assignment Center or the legacy Assignment Search index at first hand in this session (search-tool budget exhausted at the moment the Assignment Search queries ran). Consequently I can confirm that post-issuance assignment activity exists (Google Patents mirrors the Assignment Center feed), but I could not verify reel/frame numbers, execution dates, or the recorded correspondent for the entries below. I flag every unverified field explicitly rather than inventing it. Please treat the reel/frame column as a to-do against the live Assignment Center record linked at the bottom.

---

Inventors

Unusual-pattern check: Not present. There is no evidence of inventors leaving the assignee near filing. All three were Fractus personnel at filing and remained associated with the portfolio across the 2006→2023 continuations (Puente Baliarda is a Fractus co-founder and long-time CTO-level figure). No fire-sale-precursor pattern here.

Discrepancy flagged: the granted front page and USPTO bibliographic records render the second inventor as "Josep Mumbrú / Josep MUMBRU" and one IEEE citation set renders the third as "Jordi Ilaro" while the Google Patents legal-events reassignment text records the assignor as "ILARIO, JORDI." Same person, spelling variant in the recorded instrument — worth noting because a mismatch between the recorded assignor name and the printed inventor name is a common title-cleanup trigger.

---

Original assignee

Fractus, S.A. — Barcelona / Sant Cugat del Vallès, Spain. Original and current assignee; the front page lists it as both (71) Applicant and (73) Assignee.

• Primary line of business: antenna R&D and IP licensing. Fractus was founded in 1999 and did ship real product ("an active supplier of antennas… shipped millions of antennas worldwide," per its own 2009 press release and the EPO SME case study). By ~2008 it deliberately transformed its business model to patent licensing, and by the 2014–2017 period self-describes as a "leading antenna IP technology provider" with ~US$100M in licensing fees earned. The product/manufacturing line was later housed in the separate entity Fractus Antennas, S.L. (rebranded Ignion, "Virtual Antenna™"), which is not the assignee of record here.
• Current status: operating as a licensing/assertion company — not dissolved, not in bankruptcy. Fractus remains an active plaintiff as of 2024–2025 (e.g., Fractus v. Geotab, E.D. Tex., filed Dec. 6, 2024; Fractus, S.A. v. ResMed Inc. et al., No. 3:25‑cv‑02680, S.D. Cal.). Its patents are the target of Unified Patents/RPX-tracked IPRs (e.g., IPR2026‑00196, Cellco/Verizon, terminated‑settled 2026‑04‑09; IPR2026‑00191, ResMed, instituted 2026‑06‑02), though those petitions name the children (US 11,031,677 / 11,349,200 / 12,095,149), not '380 itself.

---

Assignment timeline

The Assignment Center feed as mirrored by Google Patents shows one recorded assignment event on this patent number after grant, plus the original inventor→Fractus instrument that must underlie the chain. Fields I could not verify are marked [unverified].

• YYYY‑MM‑DD [unverified] / recorded 2023‑06‑27 — Reel [unverified]/[unverified]

  • Conveyance: Assignment (Assignment of Assignors' Interest)
  • Assignor: Jordi Ilario; Josep Mumbru; Carles Puente Baliarda (recorded as "ILARIO, JORDI; MUMBRU, JOSEP; PUENTE BALIARDA, CARLES")
  • Assignee: Fractus, S.A. (Barcelona, ES)
  • Correspondent: [unverified at the Assignment Center] — the only counsel of record anywhere in this patent's file is Edell, Shapiro & Finnan, LLC, 9801 Washingtonian Blvd., Gaithersburg, MD 20878 (prosecution POA; docket 0690.0023CN-series). Whether Edell Shapiro was also the recording correspondent is not confirmed.
  • Context: confirmatory / title-perfecting assignment from the named inventors to the same original assignee — not a transfer to a third party, not a fire-sale. Re-papering inventor→company assignments on a long continuation chain (2006 priority, filed 2017, issued 2020) is a routine standing/title cleanup, and its timing (mid‑2023) sits inside Fractus's active litigation campaign.

• YYYY‑MM‑DD [unverified] / circa 2006–2007 [unverified] — Reel [unverified]/[unverified]

  • Conveyance: Assignment
  • Assignor: Puente Baliarda; Mumbru; Ilaro
  • Assignee: Fractus, S.A.
  • Correspondent: [unverified]
  • Context: original inventor-to-company assignment covering the priority application 11/614,429 (filed 2006‑12‑21) and its continuation lineage. I am inferring the existence of this instrument from the chain, not from a document I retrieved — flagged accordingly.

No other recorded instruments (no security agreements, no mergers, no change-of-name, no releases, no assignments to any LLC, aggregator, or third party) surfaced for US 10,644,380.

Note on the record I did retrieve: a legacy-assignments PDF surfaced in search showed reel/frame 024252/0234 and 025593/0823, but that document concerns an unrelated patent/party (a different assignor and correspondent) and must not be attributed to this patent.

---

Timeline diagram

timeline
    title Ownership of US 10644380
    2006 : Priority application filed by Fractus
    2017 : Continuation filed
    2020 : Patent issued to Fractus SA
    2023 : Confirmatory assignment from inventors

---

NPE / troll-pattern signals

1. Shell-entity transfer — Not present. No assignee bearing "IP / Patents / Licensing / Holdings / Ventures"; the assignee is the same Spanish operating/R&D company from filing to today. The only recorded post-grant instrument (recorded 2023‑06‑27) runs from the inventors to Fractus, i.e., toward the operating entity, not away from it.

2. Known asserter in the chain — Present (with a caveat). The assignee, Fractus, S.A., is not on the classic Acacia/Marathon/IV/Wi‑LAN list but is a well-documented high-frequency patent plaintiff and pure licensor — $100M+ licensing revenue (Fractus's own figures), 2009 multi-defendant handset suit (Samsung/LG/RIM/Pantech/Kyocera/Palm/HTC/Sharp/UTStarcom/Sanyo), 2018 carrier/CommScope actions (AT&T, T‑Mobile, Verizon), 2022 ADT/Vivint, 2024 Geotab, 2025 ResMed, and repeated IPR challenges from Unified Patents/RPX-tracked petitioners. Caveat: Fractus originates as a genuine operating company and asserts its own internally developed patents, so it is best characterized as an operating-company-turned-licensor, not a shell.

3. Repeat correspondent across the chain — Unclear / not verifiable. I could not confirm the recording correspondent for the 2023‑06‑27 entry. The only firm evidenced in the file is Edell, Shapiro & Finnan, LLC (Gaithersburg, MD), as prosecution correspondent on the continuation chain — a single-appearance fact that does not by itself establish an NPE-recording-play pattern, and per your own rule a single appearance is not a finding.

4. Cascading transfers — Not present. No series of LLC-to-LLC assignments; no two transfers within 24 months; assignee identity never changes.

5. Pre-litigation transfer — Not present / unclear. The only recorded assignment (recorded 2023‑06‑27) is >17 months before Fractus v. Geotab (filed 2024‑12‑06) and >2 years before Fractus v. ResMed (2025), so it does not fall inside the ≤6‑month window. It is a title-perfecting instrument to the already-owning entity, not a transfer arranged to set venue or standing.

6. Bankruptcy fire-sale — Not present. No Chapter 7/11 filings by Fractus; it is an active, funded licensor.

7. Privateering — Not present. There is no operating parent transferring to a proxy NPE; Fractus is the rights holder asserting its own (self-developed) portfolio.

8. Defensive aggregator — Not present. The chain does not terminate at RPX/AST/LOT/Unified/OIN; it terminates at Fractus, which is on the opposite side of the table from those organizations.

---

Verdict

NPE — moderate confidence.

Justification: the assignment chain itself is unusually clean — a single confirmatory assignment recorded 2023‑06‑27 running from the three named inventors to the same original assignee, Fractus, S.A., with no shell LLCs, no cascading transfers, no bankruptcy, and no third-party acquisition. However, the sole assignee (Fractus) has been a pure licensing and serial-assertion entity since ~2008, does not itself ship products embodying the '380 claims (its manufacturing/product line sits in the separate Fractus Antennas/Ignion entity), and is the named plaintiff in numerous suits against carriers and device makers (2009 Samsung et al.; 2018 AT&T/T-Mobile/Verizon/CommScope; 2022 ADT/Vivint; 2024 Geotab; 2025 ResMed) — that is one strong "known asserter" signal, which yields moderate rather than high confidence. No reel/frame beyond the 2023‑06‑27 entry should be cited until verified.

Verify at: https://assignmentcenter.uspto.gov/ (search by patent number 10,644,380 / application 15/856,626), mirrored at https://assignment.uspto.gov/patent/index.html — pull the reel/frame, execution date, and correspondent for the 2023‑06‑27 entry and for the original 2006/2007 inventor→Fractus instrument to close the gaps flagged above.

I'll search for the specific patent and its cited prior art references.

Let me dig deeper for the actual cited-references list for this specific patent number.

Let me try to pull the actual front-page citation list from FreePatentsOnline and Justia.

Prior-art analysis — US 10,644,380 B2

Method / limits up front. Per your instruction I searched only for the specific number 10,644,380 ("10644380"). I located: (a) the granted patent's own front page, including its partial "(56) References Cited" list (PDF at patentimages.storage.googleapis.com/b3/2f/b1/a25621b753024b/US10644380.pdf); (b) the Google Patents record (patents.google.com/patent/US10644380/en); and (c) the Justia "Patent Citations" index for 10644380. I did not obtain a complete, unabridged machine-readable citation list for this grant, and I therefore will not invent descriptions or §102 mappings for references I could not actually read. What follows is grounded in what the searches returned, with the gaps flagged.

---

1. Nature of the citation record for 10644380

10644380 is a continuation (App. 15/856,626, filed 2017‑12‑28) of 14/738,090 → 14/246,491 → 11/614,429 → priority 2006‑07‑18 (EP 06117352.2; provisionals 60/831,544 and 60/856,410). Because it is a continuation in a heavily litigated family, its Form SB/08 citation list is cumulative and enormous — the front page begins with U.S. patents back to 1963 and continues for many pages, plus a long "Other Publications" block. Key consequence for the §102 analysis:

• The critical date is 2006‑07‑18. Only art published/filed before that date is §102 prior art (pre‑AIA §102(a)/(b) governs).
• The examiner allowed the application over this art; the cited references did not produce a rejection of the issued claims. So on the face of the record, the citations functioned as background/considered art, not as anticipating references.

Partial verbatim list from the '380 front page (as returned by search; truncated):

3,079,602 A 2/1963 Du Hamel · 3,521,284 A 7/1970 Shelton · 3,599,214 A 8/1971 Altmayer · 3,622,890 A 11/1971 Fujimoto · 3,683,376 A 8/1972 Pronovost · 3,683,379 A 8/1972 Saddler · 3,689,929 A 9/1972 Moody · 3,818,490 A 6/1974 Leahy · 3,967,276 A 6/1976 Goubau · 3,969,730 A 7/1976 Fuchser · 4,021,810 A 5/1977 Urpo · 4,024,542 A 5/1977 Ikawa · 4,038,662 A 7/1977 Turner · 4,072,951 A 2/1978 Kaloi · 4,131,893 A 12/1978 Munson · 4,141,016 A 2/1979 Nelson · 4,318,109 A 3/1982 Weathers · 4,356,492 A 10/1982 Kaloi · 4,381,566 A 4/1983 Kane · 4,471,358 A 9/1984 Glasser · 4,471,493 A 9/1984 Shober · 4,504,834 A 3/1985 Garay · 4,536,725 A 8/1985 Hubler · 4,543,581 A 9/1985 Nemet · 4,571,595 A 2/1986 Phillips · 4,584,709 A 4/1986 Kneisel · 4,608,572 A 8/1986 Blakney · 4,623,894 A 11/1986 Lee · 4,628,322 A 12/1986 Marko · 4,673,948 A 6/1987 Kuo · 4,723,305 A 2/1988 Phillips · 4,730,195 A 3/1988 Phillips · 4,752,968 A 6/1988 Lindenmeier · 4,827,266 A 5/1989 Sato · 4,827,271 A 5/1989 Berneking · 4,839,660 A 6/1989 Hadzoglou · 4,843,468 A 6/1989 Drewery · 4,847,629 A 7/1989 Shimazaki · 4,849,766 A 7/1989 Inaba · 4,857,939 A 8/1989 Shimazaki · 4,860,019 A 8/1989 Jiang · 4,890,114 A 12/1989 Egashira · 4,894,663 A 1/1990 Urbish · 4,907,011 A 3/1990 Kuo · 4,912 [… list continues] …

Other Publications: Russell, D. A. et al., Dimension of strange attractors, Physical Review, Oct. 6, 1980, vol. 45, No. […]. (plus a very large NPL block)

The companion family grant US 8,738,103 (same title, same inventors, parent of '380) has an essentially identical, machine-readable citations table at uspto.report/patent/grant/8738103, and includes the Cohen and other early-2000s items listed in §2 below.

---

2. Most relevant prior art (the subset I can actually ground)

Because I could not read the full '380 citation list, I am reporting the references that are confirmed present in the family's citation record and that are substantively closest to the claimed subject matter. Descriptions marked [general knowledge] are my own technical understanding, not a quote from the document; treat them as lower confidence.

Family/reexam art worth flagging (not '380's own face list, but the same disclosure is being re‑examined): In the ex parte reexamination of the child US 11,349,200 (Reexam 90/019,275), the references cited were:

• US 2004/0137950 A1 (Bolin et al.), pub. 2004‑07‑15
• GB 2416625 A (Langley et al.), pub. 2006‑02‑01
• US 2004/0051669 A1 (Rutz et al.), pub. 2004‑03‑18 (OCR'd as "Rutzors")
• EP 1 294 049 A (Ollikainen et al.), pub. 2003‑03‑19

(ptacts.uspto.gov/.../petitions/1557927/...). These are the closest technically to the antenna‑in‑a‑handset features and are the references I'd examine first if attacking a claim of '380, even though they appear on the child's reexam record rather than '380's face.

---

3. Bottom line on §102

• I could not obtain the complete, verbatim citation list for 10644380; the list is a very long cumulative one (U.S. patents from 1963 onward + a large NPL block), carried over from the parent applications.
• No single cited reference appears to anticipate claims 1–3 (as summarized from the spec: the two‑body clamshell/slide/twist MFWD, the ≥1 GB smartphone configuration, and the camera/multimedia configuration, each combined with antenna‑contour complexity factors F21 ∈ [1.05, 1.80] and F32 ∈ [1.10, 1.90]). The cited art is antenna‑geometry art; the claims are directed to a numeric complexity‑factor window applied to a specific multifunction/multi‑body device. That combination is what distinguishes the claims, and it is absent from the references of record.
• The most plausible role for the cited art is therefore §103 obviousness background (fractal/space‑filling/self‑similar antenna geometry: Cohen patents, Puente Baliarda dissertation/IEEE, and the fractal‑dimension literature such as Russell), rather than §102 anticipation.
• Flagged uncertainty: I have not verified the complete granted claim set of 10644380 (a prior section of this analysis already noted the claim numbering of the 2nd/3rd independent claims was inferred, not confirmed). Consequently, the "which claims" column above is keyed to that reconstructed claim set, not to a verbatim grant. If you need claim‑by‑claim §102 charts, the authoritative next step is to pull the full citation list and claim set from USPTO PatentCenter for App. 15/856,626 and the '380 PDF front page directly.

Sources consulted: Google Patents US10644380; US10644380B2 PDF front page (patentimages.storage.googleapis.com/b3/2f/b1/a25621b753024b/US10644380.pdf); Justia patents.justia.com/patent/10644380; uspto.report/patent/grant/8738103; IPR2018‑01461 Ex. 1021 (ZTE v. Fractus); Reexam 90/019,275 certificate for US 11,349,200.

I'll gather the prior-art references associated with US 10,644,380 before building the § 103 analysis.

Let me pull the actual cited-art lists and any post-grant challenges naming this patent.

Obviousness Analysis — US 10,644,380 B2 under 35 U.S.C. § 103

Analyst note on sources. The "Prior Art" content surfaced for this page centers on the Inter Partes Review petition analysis for IPR2026-00191, ResMed Corp. v. Fractus, S.A. (filed Jan. 2, 2026), whose grounds are identical to IPR2025-01026, Geotab Inc. v. Fractus, S.A. (instituted Dec. 3, 2025). Those petitions attack U.S. Patent No. 11,031,677 ("the '677 patent"), the direct continuation of the patent at bar (15/856,626 → US 10,644,380 → 16/832,820 → US 11,031,677). The '677 shares the '380 specification verbatim in all material respects. That makes the petition record highly probative for '380, but not automatically transposable, for the reason explained in §VII below. Where I rely on the petition summary rather than the underlying documents, I say so.

I could not retrieve the verbatim granted claim set of US 10,644,380 (consistent with the earlier section). Claim scope below is drawn from the patent's own Summary section and family text, and I flag inference where used.

---

I. Legal framework and the effective filing date

Standard. § 103 asks whether the claimed subject matter as a whole would have been obvious to a POSITA at the time of the invention, considering (1) scope/content of the prior art, (2) differences, (3) level of ordinary skill, and (4) secondary considerations (Graham v. John Deere). KSR Int'l v. Teleflex supplies the rationales: combining known elements according to known methods to yield predictable results; substituting one known element for another to obtain predictable results; use of a known technique to improve a similar device; and design incentives/market demand.

POSITA. Reasonably defined as a person holding a bachelor's degree in electrical engineering (or physics) with 2–3 years of experience in the design of internal antennas for portable wireless devices, or equivalent. (The Board's exact formulation in the Fractus IPRs may differ; verify against the institution decisions.)

Effective filing date — decisive for art qualification. The '380 issued from an application filed Dec. 28, 2017, but claims benefit of a chain terminating in EP 06117352.2 (July 18, 2006), US 60/831,544 (July 18, 2006) and US 60/856,410 (Nov. 3, 2006), via 11/614,429 (filed Dec. 21, 2006). Because the '380 specification (including the F₂₁/F₃₂ grid methodology and the clamshell/slide/twist, ≥1 GB, ≥2 MP subject matter) is reproduced from that same disclosure, the '380 is very likely entitled to a July 18, 2006 effective filing date.

Consequence: prior art must either (a) have been publicly available before July 18, 2006 (§ 102(a)(1)), or (b) be a US patent/published application effectively filed before that date (§ 102(a)(2)). The AIA § 102/103 regime governs. This date is the single most important gate in the analysis and should be verified from the file wrapper before any ground is finalized.

---

II. The claims at issue — three independent claim families

Per the patent's Summary section, the independent claims fall into three groups, all sharing the same antenna limitation:

Dependent claims tighten the windows (e.g., F₂₁ > 1.39 & F₃₂ > 1.50; F₂₁ > 1.43 & F₃₂ > 1.63) and add thicknesses of 5–9 mm. The range-narrowing dependents are, if anything, more vulnerable than the independents, because the narrower the recited window and the closer the prior-art antenna's measured (F₂₁, F₃₂) sits to it, the stronger the In re Woodruff / In re Malagari range-optimization case.

---

III. The prior art on this page

The petition's expert calculated the complexity factors of the prior-art antennas under the '677/'380 grid methodology and obtained:

• Ciais-Quadband antenna: F₂₁ = 1.31, F₂₂ … F₃₂ = 1.57
• Nakano antenna: F₂₁ = 1.49, F₃₂ = 1.46

Both pairs fall squarely inside the '380's claimed windows (F₂₁ 1.05–1.80; F₃₂ 1.10–1.90). This is the analytical hinge of the case (see §VIII).

---

IV. Combination 1 (primary): Dou + Ciais-Quadband [+ Nakano]

What each teaches. Dou supplies a handheld wireless device architecture expressly contemplating multiple internal antennas but leaving the antenna selection to the POSITA. Ciais-Quadband supplies a compact internal PIFA covering the GSM/DCS/PCS/UMTS bands — i.e., exactly the 810–960 MHz, 1710–1990 MHz and 1900–2170 MHz regions the '380 specification identifies as its target regions. Nakano adds 2.45/5.2 GHz WLAN coverage, matching the '380's recited WiFi/Bluetooth/UWB connectivity.

Motivation (KSR).

1. Express teaching in Dou — a reference that discloses a device with "suitable internal antennas" left to the artisan is an express invitation to fill the gap; this is the strongest possible motivation showing and does not rely on hindsight.
2. Predictable results — substituting a known quadband internal PIFA into a known handset architecture to achieve the bands that architecture contemplates is a "predictable variation" of a known technique.
3. Market demand / design incentive — 2006 handsets were migrating to internal quadband antennas to support GSM/DCS/PCS/UMTS; Nakano responds to the well-documented demand for 5 GHz WLAN in handsets.
4. Art in the same field, addressing the same problem — handset antenna size/bandwidth trade-off, as acknowledged in the '380 specification itself (citing Chu and Wheeler).

Result. If the combination yields an antenna whose contour is that of Ciais-Quadband — which the petition's expert measured at (1.31, 1.57) — then all limitations of the '380 claims except the device-feature limitations of Groups A/B/C are met, and the F₂₁/F₃₂ limitations are met on the prior-art antenna as an inherent property of its geometry. A prior-art antenna's perimeter either has a given complexity under the patent's own measurement protocol or it does not; the protocol is self-executing. That is an inherency argument, not hindsight.

---

V. Combination 2: Supplying the device-feature limitations of Groups B and C

Group B (smartphone: CPU + OS running word/spreadsheet/slide apps; ≥1 GB memory). These are the classic PDA-phones of the pre-2006 era. Combine Dou + Ciais-Quadband with a handheld reference disclosing a Windows Mobile / Palm OS / Symbian-class device with an office-application suite and ≥1 GB mass storage.

• Motivation: the convergence of cellular telephony with personal-computing capability was a recognized 2006 design trend; KSR sanctions "the predictable use of prior art elements according to their established functions" — a known pocket computer with a known internal quadband antenna* is the combination of two known devices to yield an enhanced, predictable result.
• Note on the >1 GB and >2 MP thresholds: these are arbitrary capacity/specification floors. Under In re Woodruff and In re Malagari, selecting a threshold within or adjacent to a range disclosed or suggested by the art is prima facie obvious absent evidence of criticality. The '380 specification offers no criticality data for either threshold.

Group C (MMT: audio receiver; ≥2 MP image sensor/flash/zoom; ≥1 GB storage). Camera-phones of ≥2 MP and MP3-capable handsets were commercially abundant well before July 2006. The '380 specification itself concedes that the MMT is a device "including means to reproduce digital music and sound signals" and "a digital camera to record still … and/or moving images" — i.e., it admits the constituent elements as known and claims only their aggregation with the antenna. Aggregation of known, functionally independent modules without an unexpected cooperative effect is obvious.

Group A (multi-body clamshell/slide/twist). If Dou does not itself disclose a two-body mechanism, combine with any of the numerous pre-2006 clamshell or slider handset references (the hinge/slide form factors were ubiquitous). The motivation is the well-known incentive toward larger displays and compactness; the '380 specification acknowledges that flat antennas are preferred "in particular, for MFWDs which have two parts that can be shifted or twisted against each other." That admission supplies the motivation directly — an applicant's own specification can supply the motivation to combine.

---

VI. Combination 3 (secondary): Misra + Cohen (+ handset integration)

The Fractus portfolio has been repeatedly attacked on this combination. In the ZTE reexamination request (Ex. 1021 in IPR2018-01461), the requester argued Misra in view of Cohen for claims of US 7,397,431, contending that Misra discloses a multilevel radiating structure with geometric elements each associated with a frequency band, and that Cohen discloses placement of a multi-band antenna in a cellular/wireless telephone. The stated reason to combine was that both references address the same problem (small multilevel antennas) and Misra's quantitative results "bear out the predictions of Cohen."

Application here. Misra + Cohen supply (a) a multiband, geometrically rich conductive radiating structure and (b) its integration into a handheld phone. Adding a clamshell/slider handset reference and a smartphone/MMT reference supplies the device limitations. Weakness: neither Misra nor Cohen was measured for F₂₁/F₃₂ in the record I retrieved, so this combination is incomplete as to the numeric-range limitations unless an expert performs the grid measurements. It is therefore a corroborating ground, not a standalone one.

---

VII. Combination 4 (Baliarda-543): a ground that should not be transplanted from the '677

This is where the '677 petition's logic does not carry over, and I flag it explicitly because a careless analyst would import it:

• The '677 petition argued that the claims reciting a "4G communication standard" are not entitled to the 2006 priority because LTE and its bands were undefined in 2006, shifting those claims to an April 7, 2014 effective filing date. That shift renders Baliarda-543 (US 2008/0018543 A1, published Jan. 24, 2008) available as § 102(a)(2) art.
• The '380 claims as reconstructed do not recite 4G. They recite a multi-body MFWD, a smartphone with office applications and ≥1 GB memory, or an MMT with audio/camera/storage — all of which are plausibly supported by the July 18, 2006 disclosure. Absent a successful priority attack, Baliarda-543 is a publication of an application in the same family / same inventors that is not "effectively filed before" July 18, 2006 (the 11/614,429 parent was filed Dec. 21, 2006), and is therefore not prior art under § 102(a)(2).

Conclusion: the Baliarda-543 ground transfers to the '380 only if Fractus's priority claim to July 18, 2006 can be broken. I have not seen a basis for that, and I would not assert one.

---

VIII. Why the F₂₁ / F₃₂ numeric windows do not save the claims

This is the crux, and it is where the '380 is at least as exposed as the '677 — arguably more so.

1. The '380 windows are broader than the '677's. The '677 requires F₂₁ ≥ 1.20 and F₃₂ < 1.75; the '380 permits F₂₁ ∈ [1.05, 1.80] and F₃₂ ∈ [1.10, 1.90]. Both prior-art antennas measured by the petition's expert — Ciais (1.31, 1.57) and Nakano (1.49, 1.46) — satisfy the '380 windows with margin. The narrower '677 claim is the harder case; the broader '380 claim is the easier one.

2. The specification itself shows the windows capture almost the entire non-trivial design space. A perfect rectangle yields F₂₁ = 1; a Hilbert-filling curve yields F₂₁ < 2 and F₃₂ = 2. The claim's lower bounds (1.05, 1.10) exclude essentially only the bare rectangle; the upper bounds (1.80, 1.90) exclude only near-Hilbert extremes. The two worked examples in the patent — structures 1200 at (1.55, 1.58) and 1410 at (1.49, 1.85) — sit mid-window, surrounded by wide unclaimed margins. A range that excludes only the two recognized extremes is prima facie a result-effective range of a design parameter, not a criticality-bounded invention.

3. Numeric ranges are presumptively obvious. In re Woodruff; In re Malagari; KSM v. HTC (range overlap). The applicant's own argument during prosecution — quoted in the PTAB/family record — was that "the values set for complexity factors F₂₁ and F₃₂ are non-obvious" (FRAC-ADT-00000678-79). A bare assertion of non-obviousness of a numerical range, unsupported by data showing that performance varies critically at the recited boundaries, does not rebut the prima facie case.

4. Newly-named parameters of known structures are not patentable. Where a claim defines a known article by a newly devised parameter, and the prior-art article inherently possesses the recited parameter values, the claim is unpatentable. The petitioner escapes any need to show the prior-art designer knew the F₂₁/F₃₂ values — the measurement protocol is applied to the prior-art geometry, and the values fall in range. (This is precisely the structure of the petition's expert methodology.)

5. The VSWR/efficiency evidence cuts against, not for, the patentee. FIGS. 19A–19B map VSWR and efficiency against frequency, but there is no showing that antennas just inside the recited boundaries outperform antennas just outside them. Without a boundary-criticality showing, the ranges are arbitrary.

6. Corroborating clarity objection from the international phase. In the PCT family member (PCT/EP2007/006242, published as WO 2008/009391 with the A3 search report), the ISA objected under PCT Article 6 that "complexity factor" has "no well established meaning in the field of antennas"; that "one and the same antenna element could exhibit a plurality of different complexity factors depending on the definition of the grids"; and that the claims define no technical feature permitting a skilled person to design an antenna meeting his own specification. This is a clarity/enablement critique, not a § 103 ground — but it is powerful context: if the parameter is not a stable, uniquely defined technical feature, then it cannot supply the criticality that a range-claim needs to survive § 103.

---

IX. Consolidated motivation-to-combine showing

---

X. Anticipated rebuttals and how they fare

• "The prior art does not teach the F₂₁/F₃₂ ranges." Answered by inherency and mutual measurement on the same protocol. Also answered by the absence of criticality data.
• "Patch/PIFA antennas are narrowband (1–5%), unsuited to UMTS." The '380 specification makes this argument, but it is self-defeating: the specification concedes that a patch antenna is "particularly suitable for slim multifunctional devices" and that the specific complaint is bandwidth. Ciais-Quadband is a quadband PIFA — i.e., art that has already solved the bandwidth complaint the patentee raises. A reference that solves the identified deficiency cannot "teach away" from the combination (In re Fulton; In re Gurley).
• Secondary considerations. Fractus has asserted commercial success and has won jury verdicts on portfolio patents. Under WBIP the patentee must establish a nexus between the objective evidence and the claimed subject matter — specifically, that the F₂₁/F₃₂ windows (as opposed to the underlying antenna art) drove the success. Expect this to be the battleground.
• § 112 rather than § 103. The complexity-factor definitional instability (see §VIII.6) is more naturally a written-description/enablement attack. If the windows cannot be reproduced without the applicant's private grid conventions, that simultaneously weakens the § 103 rebuttal and supports § 112 invalidity.

---

XI. Status, and confidence levels

• IPR2025-01026 (Geotab v. Fractus), directed to the '677, was instituted Dec. 3, 2025 as to all challenged claims and all grounds — meaning the Board found a reasonable likelihood of prevailing on at least the Ground that met the standard. IPR2026-00191 (ResMed v. Fractus), filed Jan. 2, 2026, presents substantively identical grounds, same expert declaration (Dr. Daniel van der Weide), and seeks joinder. A related docket item lists IPR2026-00191 among proceedings referred for merits review.
• Conflict to flag with the previously generated section: that section states today's date is 2026-09-27 and describes IPR2026-00191 as "instituted June 2, 2026." This task's header states the current date is April 26, 2026, and one retrieved institution-decision document sits at a /2026/05/ URL path. These three data points cannot all be true simultaneously. I flag the inconsistency rather than reconciling it, and note that the precise institution date for IPR2026-00191 should be re-verified.
• No IPR, PGR, or reexamination directed at US 10,644,380 itself was located. The challenges are to its children (notably the '677). Because the '677 shares the '380 specification, the petition record is the best available proxy for how the '380 would be attacked, subject to the claim-scope differences in §VII.

Confidence:

• High that Combination 1 (Dou + Ciais-Quadband [+ Nakano]) is the central § 103 ground, and that the measured (F₂₁, F₃₂) values of the prior-art antennas fall within the '380's claimed windows.
• High that the numeric windows are the weak point of the claims and that the broader '380 windows are more exposed than the '677's.
• Moderate on Dou's effective filing date relative to July 18, 2006 — this must be verified, as it is dispositive of whether Dou is § 102(a)(2) art.
• Low as to the Baliarda-543 ground against '380 — I affirmatively recommend against asserting it absent a successful priority attack.
• Unverified — the verbatim granted claim language and the exact Board POSITA formulation; both should be pulled from the file wrapper and institution decisions before any ground is formalized.