US 8886954

I'll verify the bibliographic data and check for any 2026 CAFC activity before summarizing.

Let me check the CAFC docket and USPTO status specifically for the '954.

US Patent 8,886,954 B1 — Concise Summary

I searched for the exact identifier 8,886,954 and cross-checked the bibliographic data against the authoritative full text supplied in this analysis, the Google Patents/FreePatentsOnline records, and the Federal Circuit / PTAB record. I did not auto-correct any number, and I did not pull in results for neighboring patents (e.g., 8,886,945 or 8,866,954).

Bibliographic data (verified)

Uncertainty flag: The assignment reel/frame for the '954's own recording could not be confirmed from a primary USPTO source (a later section of this analysis flags this as well). Google Patents' legal-events mirror shows only "2012‑12‑12 Assigned to PROXENSE, LLC … Assignors: GIOBBI, JOHN J."

Abstract

"Systems and methods verifying a user during authentication of an integrated device. In one embodiment, the system includes an integrated device and an authentication unit. The integrated device stores biometric data of a user and a plurality of codes and other data values comprising a device ID code uniquely identifying the integrated device and a secret decryption value in a tamper proof format, and when scan data is verified by comparing the scan data to the biometric data, wirelessly sends one or more codes and other data values including the device ID code. The authentication unit receives and sends the one or more codes and the other data values to an agent for authentication, and receives an access message from the agent indicating that the agent successfully authenticated the one or more codes and other data values and allows the user to access an application."

The invention in plain language

A small, portable integrated device (a "biometric key" / "PDK" — e.g., a key fob, or the biometric subsystem built into a phone, tablet, laptop, watch, etc.) permanently stores two things in tamper-proof, write-once memory: (1) the owner's biometric template, and (2) a set of codes, principally a device ID and a secret decryption value. When a finger is scanned and the live scan matches the stored template on the device itself, the key wirelessly transmits its device ID to a nearby reader/authentication circuit. That reader does not decide on its own; it forwards the code over a network to a third-party-operated "trusted authority" (e.g., a Proxense-style registry) that checks whether the device ID belongs to a legitimately enrolled key. If the authority confirms it, an "access message" is returned and the application is unlocked (a slot machine, door lock, ATM, medical record, financial account, etc.). If the fingerprint does not match, no code is sent at all.

Layered on top: enrollment under a trusted witness/agent, age-verification registration, secure-channel setup before transmission, and an optional "code-only, no-biometric" mode.

Independent claims — plain-language overview

Claim 1 — Method (device side)

A method with these steps:

1. Persistently store, on the integrated device and in a tamper-proof, non-alterable storage element: the user's biometric data, plus a plurality of codes/data values including (a) a device ID code uniquely identifying the device and (b) a secret decryption value.
2. On a request for biometric verification, receive scan data from a biometric scan.
3. Compare the scan data to the stored biometric data for a match.
4. On a match, wirelessly send one or more codes — including the device ID — "for authentication to a third party that operates a trusted authority."
5. Receive, at an application, an access message from that trusted authority indicating successful authentication, and thereby allow the user access to the application.

Claim 12 — Integrated device (apparatus)

Same substance as claim 1, as hardware: a persistent storage media holding the biometrics + device ID + secret decryption value in non-alterable form; a validation module that receives biometric scan data, compares it, and sends the code(s) for authentication by the third-party-operated trusted authority; and an RF communication module that receives the access message and permits access to an application.

Claim 16 — Method (reader/verifier side)

A method (executed by a programmed processor) to authenticate an already biometrically-verified user:

1. Wirelessly receive the code(s)/data values including the device ID of the integrated device associated with a biometrically verified user, the device ID having been registered with a third-party-operated trusted authority.
2. Request authentication of the device by the trusted authority using those codes.
3. Receive, at an application, an access message from the trusted authority indicating successful authentication.
4. In response, allow the biometrically verified user access to the application.

Claim 22 — System

An integrated hardware device (stores biometrics + device ID + secret decryption value in non-alterable storage; wirelessly sends the code(s) including the device ID) plus an authentication circuit that receives the codes, sends them to the third-party-operated trusted authority, receives the access message, and allows the user to access the application.

Selected dependents: age/DOB verification registered with the code (3, 14, 17); codes transmitted over a network (2, 13); code indicates biometric verification succeeded (4); enumerated biometrics — fingerprint, palm, retinal, iris, hand geometry, face, signature, voice (5, 26); device forms — phone, tablet, laptop, MP3 player, gaming device, watch, key fob (6, 25); application types — casino machine, keyless lock, garage door opener, ATM, hard drive, software, web site, file, financial account (7–9, 19–21, 27–29); secure channel before sending (10, 18); code-only mode without a biometric request (11); LED to prompt the scan (15); device requests a scan on an authentication request (23); no code sent if verification fails (24).

2026 litigation / PTAB posture (as of April 26, 2026)

• Live appeal: Proxense, LLC v. Apple Inc., Fed. Cir. No. 25-2168, filed Sept. 30, 2025; originating IPRs 2024-00233, 2024-01334, and 2024-00846. Proxense's opening brief was filed April 3, 2026 (captioned in the brief as Proxense v. Apple, Microsoft), arguing the PTAB violated the APA by staying a parallel ex parte reexamination (90/015,052) that was trending toward patentability, and disputing the construction of "a third party that operates a trusted authority." Outcome per RPX Empower: pending, no disposition summary.
• Second CAFC docket, No. 26-1283, is listed in the Google Patents litigation panel for the '954. I could not independently verify its parties, originating IPR, or status in this pass — flagged as unverified, consistent with the earlier section.
• Claim outcomes (refining the earlier PTAB section): the two merits FWDs, taken together, cover all 29 claims. IPR2024-00233/01334 held claims 1–7, 10, 12–19, 22–27 unpatentable; the IPR2024-00846 FWD (as recited verbatim in Proxense's own April 3, 2026 appellate brief) determined claims 1, 2, 4, 5, 7–13, 15, 16, 18–24, and 26–29 unpatentable. The union of the two is claims 1–29.
  • ⚠️ Contradiction with the earlier "PTAB challenges" section: that section stated claims 8, 9, 11, 20, 21, 28, 29 were "never reached"/"untested" because Apple's third petition (IPR2024-01485, Burger/Robinson/Orsini) was denied. On the '846 FWD as recited in Proxense's brief, those dependent claims were reached and held unpatentable in the Microsoft IPR. The "untested dependent claims" characterization appears incorrect or, at minimum, superseded — treat the '846 FWD itself as the authority.
• The patent expired Dec. 20, 2025, so any remaining assertion is for past damages only (no injunctive relief).

Confidence and caveats

• High confidence in title, inventor, assignee, application number, filing date (Dec. 10, 2012), issue date (Nov. 11, 2014), priority date (Dec. 20, 2004), abstract, all 29 claims, and the content of the four independent claims — these come from the authoritative full text in this analysis.
• Moderate confidence on the current CAFC status: 25-2168 and the April 3, 2026 opening brief are well-sourced (RPX Empower; ip fray; the brief PDF itself). 26-1283 remains unverified.
• Corrected vs. prior section: the "untested dependent claims 8, 9, 11, 20, 21, 28, 29" conclusion is contradicted by the '846 FWD as recited in Proxense's brief (see flag above).
• I did not fabricate any reel/frame, docket number, or party attribution; where a value could not be confirmed from a primary source, I have said so rather than supplying one.

I'll research the litigation history for US Patent 8886954. Let me run several targeted searches.

Let me dig deeper into the specific district court cases and their outcomes.

Litigation Known to Involve U.S. Patent No. 8,886,954

Important framing note: I interpreted the identifier literally as 8,886,954 ("Biometric Personal Data Key (PDK) Authentication," Proxense, LLC, granted Nov. 11, 2014, application 13/710,109, priority Dec. 20, 2004). All results below are for that exact patent. Where a docket appears in the same campaign but I could not confirm the '954 patent itself was asserted, I say so explicitly rather than implying it.

I was working from the patent's Google Patents page (litigation links) plus docket aggregators (Docket Alarm, CourtListener, Justia, RPX Empower, Unified Patents portal, ip fray) and PTAB filings. Some petitioner attributions to specific IPR numbers could not be fully cross-verified and are flagged.

A. District Court Litigation (Patent '954 asserted)

1. Proxense, LLC v. Microsoft Corporation

• Plaintiff: Proxense, LLC
• Defendant: Microsoft Corporation
• Jurisdiction: U.S. District Court, Western District of Texas (Waco Division), Judge Alan D. Albright
• Case No.: 6:23-cv-00319 (‑ADA)
• Filing date: May 2, 2023 (the AO 120 form initially listed 4/28/2023; the clerk noted the date did not reflect filing and an amended form was to be filed)
• '954 in suit: Yes — attached as "Exhibit 2 – U.S. Patent No. 8,886,954," alongside U.S. 8,352,730; 9,298,905; 8,646,042; 9,679,289; 10,073,960
• Status/Outcome: Microsoft moved to dismiss for failure to state a claim; Proxense opposed. The case is stayed pending the parallel IPR(s). No merits judgment on '954.
• Sources: https://www.courtlistener.com/docket/67312721/proxense-llc-v-microsoft-corporation/ ; Google Patents family/litigation listing

2. Proxense, LLC v. Google LLC and Google Payment Corp.

• Plaintiff: Proxense, LLC
• Defendants: Google LLC and Google Payment Corp.
• Jurisdiction: W.D. Tex. (Waco Division), Judge Alan D. Albright
• Case No.: 6:23-cv-00320 (‑ADA)
• Filing date: May 2, 2023
• '954 in suit: Yes (asserted family included U.S. 10,073,960; 8,352,730; 8,646,042; 8,886,954; 9,298,905; 9,679,289)
• Status/Outcome: Settled and dismissed with prejudice. A joint stipulation of dismissal (Fed. R. Civ. P. 41(a)(1)(A)(ii)) was filed Jan. 13, 2025, dismissing Proxense's infringement claims with prejudice (defendant's claims/defenses without prejudice); the record indicates Google took a license. One aggregator lists the case as closed Aug. 4, 2025.
• Sources: ip fray, https://ipfray.com/staging1/google-follows-samsung-in-settling-patent-dispute-with-biometrics-firm-proxense-in-western-district-of-texas-apple-microsoft-intel-lg-cases-still-pending/ ; PTAB termination order in IPR2024-00233 (Google/Proxense joint motion)

3. Proxense, LLC v. Apple Inc.

• Plaintiff: Proxense, LLC
• Defendant: Apple Inc.
• Jurisdiction: W.D. Tex. (Waco Division), Judge Alan D. Albright
• Case No.: 6:24-cv-00143 (‑ADA)
• Filing date: March 18, 2024
• '954 in suit: Yes (with U.S. 10,698,989; 8,352,730; 8,646,042; 9,048,188; 9,298,905)
• Status/Outcome: Stayed (pending the parallel IPRs). Proxense later moved to amend to add the '289 patent. Apple filed three IPRs against '954. Per the PTAB's final written decisions, the challenged '954 claims were held unpatentable; Proxense has appealed. No district-court merits judgment.
• Sources: https://www.docketalarm.com/cases/PTAB/IPR2024-01485/Apple_Inc/docs/.../Ex_1036_Scheduling_Order.pdf ; https://litigation.rpxcorp.com/federal_circuit/[3015503](/patent/3015503)-proxense-v-apple

B. Court of Appeals for the Federal Circuit

4. Proxense, LLC v. Apple Inc.

• Appellant: Proxense, LLC; Appellee: Apple Inc.
• Court: U.S. Court of Appeals for the Federal Circuit
• Case No.: 25-2168
• Filed: Sept. 30, 2025 (appeal received 9/26/2025)
• Originating PTAB cases (per RPX): IPR2024-00846, IPR2024-01334, IPR2024-00233
• Status: Pending. This is Proxense's appeal of the PTAB's unpatentability determinations on '954. Proxense's appellate briefing argues the PTAB violated the APA by staying a parallel ex parte reexamination that was on the verge of confirming patentability, and disputes the PTAB's construction of "a third party that operates a trusted authority."
• Sources: https://dockets.justia.com/docket/circuit-courts/cafc/25-2168 ; ip fray, https://ipfray.com/ptab-abused-discretion-by-staying-reexamination-that-favored-patentee-proxense-tells-federal-circuit/

5. Federal Circuit Case No. 26-1283

• The Google Patents litigation panel for '954 lists a Court of Appeals for the Federal Circuit case, No. 26-1283, filed by Proxense (the other CAFC entry in the family). I could not confirm the appellee or the exact originating proceeding with high confidence; it is most plausibly the Microsoft-side appeal of the '954 IPRs, but I am flagging that as unverified.
• Source: https://patents.google.com/patent/[US8886954](/patent/US8886954)/en (litigation panel)

C. PTAB (Inter Partes Review) Proceedings on '954

• Note: Petitioner attribution comes from the Docket Alarm / PatentPia PTAB listings. The IPR2024-00233 termination order caption reads "GOOGLE LLC and APPLE INC., Petitioner v. PROXENSE, LLC," and ip fray states the appeal concerns "two final written determinations," which does not map cleanly onto the three FWD-designated numbers. Treat the petitioner↔number mapping for 00233/00846/01334/01327 as high-confidence but not fully reconciled.
• There is also a parallel ex parte reexamination at the USPTO Central Reexamination Unit (CRU), which the PTAB stayed — the subject of the CAFC appeal above.
• Sources: https://portal.unifiedpatents.com/ptab/case/IPR2024-00233 ; .../IPR2024-00846 ; .../IPR2024-01327 ; .../IPR2024-01334 ; .../IPR2024-01485 ; https://www.docketalarm.com/search/?q=patent%3A%288886954%29 ; PTAB termination order in IPR2024-00782 et al. (Google/Proxense)

D. Same-Campaign Cases Where '954 Does NOT Appear to Have Been Asserted

To avoid over-inclusion, I note these related Proxense suits that surfaced in the same enforcement campaign but whose asserted-patent lists did not include 8,886,954:

• Proxense, LLC v. Samsung Electronics Co., Ltd. et al. — W.D. Tex., No. 6:21-cv-00210, filed Mar. 5, 2021. Reported asserted patents: U.S. 8,352,730; 10,698,989; 9,049,188; 9,235,700; 9,298,905. Settled on the eve of trial (2023), consent judgment/dismissal with prejudice. '954 is not in the reported list.
• Proxense, LLC v. Intel Corp. — W.D. Tex., No. 6:24-cv-00283, filed May 23, 2024. Reported patents: U.S. 8,036,152; 8,219,129; 8,352,730; 8,457,672; 9,265,043. ip fray reports a settlement (Mar. 24, 2026). '954 is not in the reported list.
• Proxense, LLC v. [LG Electronics Inc.](/litigations/by-plaintiff/LG%20Electronics%20Inc.) et al. — W.D. Tex., No. 6:24-cv-00302, filed May 31, 2024. Asserted-patent list (with respect to '954) could not be confirmed from the sources retrieved.
• Proxense, LLC v. Hyundai Motor Company et al. — W.D. Tex., No. 6:25-cv-00016, filed Jan. 14, 2025 (Hyundai Motor Co., Hyundai Motor America, Genesis Motor America, Kia Corp., Kia America). '954 assertion not confirmed.
• Sources: https://www.patsnap.com/de/resources/blog/litigation/proxense-v-samsung-biometric-pdk-authentication-patents-patsnap/ ; https://goldencompass.patentpia.com/en/report/reportResult.hs?... ; https://dockets.justia.com/docket/texas/txwdce/6:2025cv00016/1172820476 ; https://dockets.justia.com/docket/texas/txwdce/6:2024cv00302/1172790008 ; ip fray (Intel settlement)

E. Bottom Line

• Yes, litigation is known: '954 has been asserted by Proxense, LLC in the Western District of Texas against Microsoft (6:23-cv-00319), Google (6:23-cv-00320), and Apple (6:24-cv-00143) — all before Judge Alan D. Albright.
• Outcomes to date: Google settled/dismissed with prejudice (Jan. 13, 2025); Apple and Microsoft cases are stayed pending IPRs. No district-court merits ruling on '954.
• PTAB: Five IPRs on '954 (00233, 00846, 01327, 01334, 01485); three are reported as final written decisions (adverse to Proxense), two as not instituted. The adverse decisions are on appeal at the Federal Circuit (25-2168, and 26-1283 listed by Google Patents).
• Caveats: (1) Petitioner-to-IPR-number attributions are drawn from aggregators and are high-confidence but not fully cross-verified; (2) I could not confirm the appellee/scope of CAFC No. 26-1283; (3) I did not auto-correct any patent or case numbers, and the two dates that surface anomalously (the AO 120 "4/28/2023" vs. actual filing 5/2/2023 for the Microsoft case) are reported as they appear in the record.

I'll research the PTAB proceedings listed in the patent record and cross-check against the USPTO ODP block.

Let me dig into the specific decisions, claim outcomes, and appeals.

PTAB Analysis — U.S. Patent No. 8,886,6954… i.e., US 8,886,954 B1 ("Biometric Personal Data Key (PDK) Authentication", Proxense LLC)

⚠️ Data-integrity flag up front. The structured "PTAB proceedings on file" block states that the USPTO ODP API returns no AIA trial proceedings for this patent. That block is wrong or badly stale for US 8,886,954. The authoritative patent record in this prompt (Google Patents / USPTO data) lists five AIA trials, and live web search confirms all five with docket-level detail. I therefore treat the ODP "none" as a data gap, not as a substantive finding, and report the five proceedings surfaced by the patent record and verified by web search. Do not rely on the ODP block for this patent.

Proceedings overview

Five AIA proceedings on file: two reached Final Written Decisions (one invalidating the independent claims; one appealed), two were denied institution, and one (Apple's second petition, IPR2024-01485) was denied institution as an improper follow-on petition. Bottom line for a defendant: the four independent claims (1, 12, 16, 22) have been held unpatentable, and every claim that was not reached (8, 9, 11, 20, 21, 28, 29) is a dependent claim hanging off a canceled independent — so the patent presents essentially no live claim to assert, subject only to the pending Federal Circuit appeal. Combined with the patent's anticipated expiration on 2025-12-20, this is a patent in a very weak defensive posture for the owner and a strong one for any target.

IPR2024-00233 — Google LLC v. Proxense, LLC (joined with IPR2024-01334, Apple Inc.)

• Type: Inter Partes Review
• Filed: 2024-01-17
• Status: Final Written Decision (per structured record); FWD at Paper 35 (the "'233 FWD"); the joined Apple proceeding (IPR2024-01334) is docketed "Final Written Decision – Appealed," terminated 2025-06-17.
• Judge panel: The related Apple panel is David C. McKone, Kevin F. Turner, Thu A. Dang, and Norman H. Beamer (per IPR2024-01334 docket). McKone is confirmed on the merits at the April 2025 oral hearing transcript for the companion '233 matter. Treat the "233 panel as the same core group (Dang/Turner/McKone) — I could not confirm the exact three-judge roster for '233 from a primary source.
• Petition grounds: Pre-AIA 35 U.S.C. § 103. Ground 1: claims 1, 2, 4–7, 10, 12, 13, 15, 16, 18, 19, and 22–27 obvious over Ludtke (U.S. Pat. No. 7,188,110); dependent claims 3, 14, 17 obvious over Ludtke in view of Kon (U.S. Pub. No. 2002/0046336). (Apple's IPR2024-01334 challenged a subset on the same Ludtke/Kon art and was joined.)
• Institution decision: Instituted (the merits proceeded to a hearing on 2025-04-22). The Board construed "a third party that operates a trusted authority" as "a trusted authority that is an entity separate from the parties to a transaction," and expressly held that "such a transaction is not limited to those in which the application being accessed is a party" — a construction that defeated Proxense's core non-infringement/validity argument.
• Final Written Decision: All challenged claims held unpatentable. Specifically: claims 1, 2, 4–7, 10, 12, 13, 15, 16, 18, 19, and 22–27 unpatentable under pre-AIA § 103 over Ludtke; dependent claims 3, 14, and 17 unpatentable over Ludtke in view of Kon. The panel dismissed Proxense's parallel-rexam argument (the CRU's 90/015,052 determination on the '730 patent) as "not an assessment on the merits" and "of no relevance."
• Settlement / termination: Google reportedly settled its Proxense dispute (reported 2025-01-14) and dismissed its challenge; Apple remained as petitioner in the joined proceeding. Terms are confidential.
• Appeal: Yes. Consolidated Federal Circuit appeal 25-2168 (Proxense, LLC v. Apple Inc.), filed 2025-09-30, originating from IPR2024-00233, IPR2024-01334, and IPR2024-00846. Proxense's opening brief was filed 2026-04-03; a further appeal (26-1283) is also on the CAFC docket per the structured record.
• Post-decision motion: Proxense filed a Request for Director Review (~July 2025), arguing the FWD violated the APA (Vivint) by reaching a result inconsistent with the CRU's 90/015,052 determination on substantially identical '730 claims, without a reasoned explanation. Outcome not confirmed in my sources.
• Defensive value: The load-bearing independent claim 1 is dead (and claims 12, 16, 22 with it). Any infringement theory built on the independent claims is now fought on invalidated ground. This is the single most important proceeding on the patent.

IPR2024-00846 — Microsoft Corporation v. Proxense, LLC

• Type: Inter Partes Review
• Filed: 2024-04-26 (DocketAlarm); instituted in 2024 (referenced as "currently pending" in Proxense's 2024-09-27 mandatory notices).
• Status: Final Written Decision – Appealed; terminated 2025-10-17.
• Judge panel: Thu A. Dang, Kevin F. Turner, David C. McKone (docket also lists Alan D. Albright and Kalyan K. Deshpande across the case history).
• Petition grounds: § 103 challenge to the '954 claims as reported in the parallel Microsoft petition — claims 1–7, 10, 12–19, and 22–27. (Treat the exact art combination as unconfirmed; Microsoft used different references than the Google/Apple Ludtke case.)
• Institution decision: Instituted (the institution is the stated premise for denying Microsoft's conditional joinder petition in IPR2024-01327).
• Final Written Decision: Issued 2025-10-17. I could not confirm the claim-by-claim verdict from a primary source. It is one of the two FWDs consolidated into the Federal Circuit appeal described in Proxense's 2026-04-03 opening brief, so at minimum the FWD resolved the merits against Proxense on the challenged claims — but do not quote a specific claim disposition for '846 without pulling the FWD itself.
• Settlement / termination: None reported; the case ran to FWD.
• Appeal: Yes — consolidated into 25-2168 (and the later 26-1283).
• Defensive value: A second, independent petitioner (Microsoft) took the patent to a final merits loss for the owner. Even a defendant who is not a Microsoft privy gets the benefit of the resulting public record.

IPR2024-01327 — Microsoft Corporation v. Proxense, LLC

• Type: Inter Partes Review (with a Conditional Motion for Joinder)
• Filed: 2024-08-20
• Status: Not Instituted – Procedural (institution denied).
• Judge panel: Thu A. Dang, Kevin F. Turner, David C. McKone.
• Petition grounds: § 103 against claims 1–7, 10, 12–19, and 22–27.
• Institution decision: Denied 2024-11-20. Microsoft expressly conditioned its petition and joinder motion on the Board having denied institution in IPR2024-00846. Because the Board instituted '846, the condition failed and the Board denied both the petition and the conditional joinder. (Not a merits ruling.)
• Final Written Decision: None (never instituted).
• Settlement / termination: N/A.
• Appeal: None.
• Defensive value: Procedural only. It signals Microsoft was hedging against a possible '846 denial; it produced no estoppel and no claim rulings.

IPR2024-01334 — Apple Inc. v. Proxense, LLC

• Type: Inter Partes Review
• Filed: 2024-08-22
• Status: Final Written Decision – Appealed; terminated 2025-06-17.
• Judge panel: David C. McKone, Kevin F. Turner, Thu A. Dang, and Norman H. Beamer.
• Petition grounds: § 103 over Ludtke and Kon (a subset of claims — the same art family the Board accepted in '233).
• Institution decision: Instituted and joined with IPR2024-00233; the joined proceeding produced the single FWD captioned "IPR2024-00233 (joined with IPR2024-01334)."
• Final Written Decision: Same disposition as the joined '233 FWD — claims 1, 2, 4–7, 10, 12, 13, 15, 16, 18, 19, 22–27 unpatentable over Ludtke; claims 3, 14, 17 over Ludtke+Kon.
• Settlement / termination: None reported for Apple.
• Appeal: Yes — consolidated into 25-2168.
• Defensive value: Apple's petition is what kept the invalidating FWD alive after Google exited; it is the vehicle behind the Apple-related CAFC appeal.

IPR2024-01485 — Apple Inc. v. Proxense, LLC

• Type: Inter Partes Review
• Filed: 2024-10-16
• Status: Not Instituted – Procedural (institution denied).
• Judge panel: McKone, joined by Dang and Turner (Paper 10).
• Petition grounds: § 103 against all 29 claims of the '954 patent, over Burger, Robinson, and Orsini — a different art set from the Google/Apple Ludtke case.
• Institution decision: Denied 2025-02-26. Applying the Board's multiple-petition guidance, the panel held that "different invalidity grounds and challenging more claims are insufficient justifications for instituting a later-filed petition when the first petition was already instituted." Apple's only added claims were dependents (8, 9, 11, 20, 21, 28, 29) not asserted in the co-pending litigation, and Proxense would not stipulate to a covenant not to sue on them.
• Final Written Decision: None.
• Settlement / termination: N/A.
• Appeal: None.
• Defensive value: Important gap-filler. The Burger/Robinson/Orsini art on claims 8, 9, 11, 20, 21, 28, 29 was never tested on the merits. Those claims remain formally untested — but see the strategic note below on why that offers the patent owner little comfort.

Strategic summary

Which claims are canceled vs. sustained vs. untested. The two merits proceedings (IPR2024-00233/01334 joined, and IPR2024-00846) put all four independent claims — 1 (method), 12 (integrated device), 16 (method), and 22 (system) — on the unpatentable side, along with their challenged dependents (2–7, 10, 12–19, 22–27 as applicable per the '233/01334 FWD). The claims that were never reached are 8, 9, and 11 (depending from claim 1), 20 and 21 (depending from claim 16), and 28 and 29 (depending from claim 22). Because § 112/claim-dependency law makes a dependent claim incorporate its parent, the "surviving" dependent claims cannot stand on their own once independents 1, 16, and 22 are canceled — so there is no meaningful live claim left to assert. Practically, the patent is a shell, subject only to the Federal Circuit. (Note also the patent's record shows anticipated expiration on 2025-12-20; the term has now run.)

Estoppel landscape. IPR estoppel under 35 U.S.C. § 315(e)(2) now binds the petitioners who reached FWD — Google, Apple, and Microsoft — and their privies/real parties-in-interest, as to any ground they raised or reasonably could have raised in the instituted proceedings (00233/01334 and 00846). For a new, unprivied defendant, those grounds are not estopped: the Ludtke and Kon combinations, and the Burger/Robinson/Orsini combination on the untested dependents, remain available. But practically, the winning Ludtke argument is public, road-tested, and embodied in two FWDs — it is the obvious template for any new § 103 challenge, and the patent owner has already litigated and lost that fight.

Pattern signals. A defensive-aggregator/NPE-defense dynamic is clearly present: the '954 patent was attacked by three separate large-tech defendants (Google, Apple, Microsoft) in a tightly coordinated multi-petition campaign (three parallel IPRs per patent family in 2024, plus three ex parte reexaminations 90/015,052–054 on family members). Unified Patents appears in the litigation graph as a petitioner/aggregator source. The same petitioner (Apple) filed two IPRs on this patent ('1334 and '1485), and the Board shut down the second as an improper follow-on. The patent owner (Proxense, via Hecht Partners) has fought hard — director review request and a consolidated CAFC appeal — but the family record shows Proxense also took adverse judgment canceling all claims of sibling U.S. Pat. No. 8,352,730 in IPR2024-00775/'232 (2025-04-28), which is a strong negative signal for the whole family.

Recommended next steps

• Defendant with a demand letter citing US 8,886,954: lead with the FWD. The Board determined the independent claims unpatentable — quote the disposition: claims 1, 2, 4–7, 10, 12, 13, 15, 16, 18, 19, and 22–27 unpatentable under pre-AIA § 103 over Ludtke; claims 3, 14, and 17 unpatentable over Ludtke in view of Kon. Cite the panel's construction of "a third party that operates a trusted authority" and its holdings that the CRU's contrary '730 reexam determination is "not an assessment on the merits" and "of no relevance."
  • Primary sources to link: the PTAB dockets at https://portal.unifiedpatents.com/ptab/case/IPR2024-00233, https://portal.unifiedpatents.com/ptab/case/IPR2024-01334, https://portal.unifiedpatents.com/ptab/case/IPR2024-00846, and the CAFC appeal at https://litigation.rpxcorp.com/federal_circuit/[3015503](/patent/3015503)-proxense-v-apple. The '233 FWD is Paper 35; the '1485 denial is Paper 10 (2025-02-26); the '1327 denial is the 2024-11-20 institution decision.
• Pull the primary documents I could not fully verify: (1) the IPR2024-00846 FWD (2025-10-17) — confirm the exact claim-level verdict; (2) the '233 FWD (Paper 35) for verbatim claim numbers and reasoning; (3) the director-review disposition on the '233 FWD; and (4) the CAFC 26-1283 docket and the status of consolidated appeal 25-2168.
• Trial-stage milestones: no instituted IPR on this patent is currently active — both live challenges concluded at FWD, and both new petitions ('1327, '1485) were denied. The only live front is the Federal Circuit appeal (25-2168 / 26-1283), where Proxense is the appellant. There is no PTAB 1-year clock running.
• "No PTAB activity" caveat: the ODP block's "none" is false for this patent — flag it internally. Five AIA proceedings exist and are documented on the public PTAB dockets. (For completeness, the same ODP/record set shows no granted CBM or PGR on the '954 — the AIA activity here is all IPR.)

I'll search for the assignment record and ownership history for this patent.

Key finding located. Let me pin down the reel/frame for the '954 patent itself and check for any additional assignments.

I've gathered enough to reconstruct the record. Note upfront: USPTO Assignment Center is not directly machine-queryable from here, so the reel/frame below comes from an authoritative USPTO "Patent Assignment Abstract of Title" excerpt surfaced inside a PTAB exhibit, plus Google Patents' legal‑events mirror. Where a reel/frame could not be confirmed for the '954 itself, I say so rather than invent one.

Inventors

• Sole named inventor on US 8,886,954. The parent application (US 11/314,199, filed 2005-12-20, now US 8,352,730) and the provisional priority documents (60/637,538; 60/652,765) name only Giobbi.
• Contrast with sibling Proxense patents: other family members (e.g., US 10,374,795; US 10,373,795-line; the healthcare/tracking patents) name co-inventors David L. Brown (Jupiter, FL), Fred S. Hirt (Brookfield, IL), Ryan Gallivan, Kent Yundt, Bryan Kelly. The '954 is a single-inventor patent, which is consistent with it being an early core disclosure rather than a later collaborative filing.
• No "inventor flight" pattern. Giobbi did not depart the assignee; he is the assignee's principal and remained CEO of Proxense as of an April 2026 interview with ip fray. There is no evidence of inventors leaving within 12 months of filing to precede a portfolio fire-sale.

Original assignee

• Proxense, LLC (a.k.a. "Proxense"), Bend, OR — recorded address in the assignment record: 689 NW Stonepine Drive, Bend, Oregon 97701.
• The patent was prosecuted and issued in Proxense's name (the printed patent's applicant field reads "Proxense, LLC, Bend, OR"), i.e., Proxense was the owner-applicant, not a downstream transferee.
• Primary line of business: biometric/"Personal Data Key" (PDK) authentication, proximity detection, secure mobile data/exchange and mobile-payment technology. Prosecution/record attorney firm of record across the portfolio: Patent Law Works LLP (Salt Lake City, UT).
• Did it ship a product embodying the claims? Proxense was a genuine operating startup from ~2001/2004–2012 ("employing over thirty engineers"), developing mobile-payment and PDK hardware. It did not achieve commercial scale; third-party databases characterize it as a former operating company and a failed startup (Stanford NPE litigation database; ip fray). It has no current product in commerce.
• Current status: Operating as a patent-enforcement entity — publicly identified as an NPE by Unified Patents ("owned and asserted by Proxense LLC, an NPE"), by the Stanford NPE Database, and by ip fray ("a non-practising entity (NPE) and former operating company"). No bankruptcy, dissolution, or acquired-by-third-party event appears in the record.

Assignment timeline

The '954 has one recorded assignment — there is no post-issuance transfer chain. (I also surface the sibling-family recording because it is the only reel/frame actually retrievable and it reveals the correspondent of record.)

• Executed on or about 2012-12-10/12 / recorded 2012-12-12 — Reel/frame not independently confirmed (Google Patents legal-events entry: "2012-12-12 Assigned to PROXENSE, LLC … Assignors: GIOBBI, JOHN J. … ASSIGNMENT OF ASSIGNORS INTEREST")

  • Conveyance: Assignment
  • Assignor: Giobbi, John J.
  • Assignee: Proxense, LLC
  • Correspondent: not confirmed for this recording
  • Context: Founder-to-company assignment (confirmation of ownership) executed essentially contemporaneously with the 2012-12-10 continuation filing — an internal ownership-perfection step, not a transfer-to-asserter.

• Family/sibling data point — NOT a record on the '954: For application 14/521,982 (= US 9,298,905, a later continuation in the same family), the USPTO Patent Assignment Abstract of Title states Total Assignments: 1:

  • Received/recorded 2015-01-29 — Reel 034843 / Frame 0145
  • Conveyance: Assignment of assignors' interest
  • Assignor: Giobbi, John J.
  • Assignee: Proxense, LLC, 689 NW Stonepine Drive, Bend, Oregon 97701
  • Correspondent: Paige M. Alsbury, 201 S. Main Street, Suite 250, Salt Lake City, UT 84111
  • Context: Founder-to-company assignment (ownership confirmation for that continuation). Source: USPTO Patent Assignment Abstract of Title reproduced in PTAB Ex. 1022-2, IPR2024-00846.
  • Flag: This is the only correspondent name recoverable for the whole family, and it aligns with Proxense's regular prosecution firm (Patent Law Works LLP, Salt Lake City) — i.e., an ordinary patent-prosecution correspondent, not a repeat NPE-recording lawyer. It does not recur across multiple assignees because there are no multiple assignees.

If the '954 itself carries a reel/frame, I could not retrieve it. Google Patents' legal-events mirror shows only the date (2012-12-12), not a reel. I state the sibling reel (034843/0145) expressly as a family fact and do not attribute it to the '954.

Timeline diagram

timeline
    title Ownership of US 8886954
    2004 : Priority date Dec 20
    2005 : Parent application filed
    2012 : Continuation filed Dec 10
         : Giobbi assigns to Proxense LLC
    2014 : Patent issued Nov 11
    2021 : First suit filed against Samsung
    2023 : Suits vs Google and Microsoft
    2024 : Suits vs Apple and Intel
    2025 : Patent expires Dec 20
    2026 : CAFC appeal still pending

NPE / troll-pattern signals

Verdict

NPE — moderate confidence.

The assignment record itself is unusually clean for an NPE story: US 8,886,954 has a single recorded assignment — inventor Giobbi → Proxense, LLC (recorded 2012-12-12) — and no shell-entity hop, no cascading LLCs, no pre-litigation transfer, and no bankruptcy. The NPE call therefore does not rest on chain-based tell-tales (all six of signals 1, 4, 5, 6, 7, 8 are absent); it rests on one strong signal (#2): the sole owner/assignee, Proxense, LLC, is a documented non-practicing asserter (Unified Patents "an NPE"; Stanford NPE Database; ip fray), which has sued Google, Apple, Microsoft, Samsung, Intel, LG and Hyundai while shipping no product. One strong signal maps to moderate confidence under the rubric; a reviewer who also weights the documented "former operating company / no product in commerce" status could argue high. No reel/frame is asserted for the '954's own recording because none could be confirmed.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent 8886954; mirror: https://assignment.uspto.gov/patent/index.html).

Cross-check against prior sections: No contradictions found. The prior summary's "assignment recorded 2012-12-12," "continuation of US 11/314,199," and "expired — lifetime / anticipated expiration Dec. 20, 2025" all reconcile with the assignment data above. One nuance worth carrying forward: because the patent expired 2025-12-20 yet the CAFC appeals remain live in 2026, any continued assertion is necessarily for past damages, not injunctive relief.

I'll research this systematically, starting with the patent's prior-art citations and the most likely §102 references.

Let me research the key IPR prior-art references and additional §102 candidates.

I've gathered substantial material. I hit the step limit before completing every lookup, so I'll flag confidence levels explicitly and note where I could not verify a citation. Everything below is grounded either in the authoritative patent text (its front‑page "References Cited" list) or in the searches completed.

Prior‑Art Analysis: US 8,886,954 B1 ("Biometric PDK Authentication")

Method note / limitations

• The '954 patent lists ~158 cited references on its face (patents, published applications, and foreign/WO documents), plus an "Other References" section. I could not individually retrieve and read all 158; I prioritized the references that are structurally closest to the independent claims (local biometric match → wireless release of a code → external authentication), because those are the only realistic §102 candidates.
• §102 vs §103 caution: The independent claims (1, 12, 16, 22) require several elements that the closest references do not all disclose in a single document — notably (i) a third party that operates a trusted authority, (ii) the application itself receiving the access message, (iii) a secret decryption value stored alongside the device ID, and (iv) storage "not capable of being subsequently altered" (write‑once). A true §102 anticipation is therefore unlikely for most references; these are better characterized as §103 primary references. I say so explicitly per reference rather than overstating anticipation.

Tier 1 — Closest references (single‑reference §102 candidates for claims 1/12/16/22)

1. US 6,041,410 A — "Personal identification fob"

• Inventors/Assignee: Hsu, Shi‑Ping; Ling, James M.; Messenger, Arthur F.; Evans, Bruce W. — TRW Inc. (Redondo Beach, CA). Also published as EP 0 924 656 A2/A3 (app. 98‑123023.8).
• Filing date: Dec. 22, 1997 (priority US 995,267); Issued: March 21, 2000.
• Description (verified): A handheld fob contains a fingerprint sensor, a memory storing a reference fingerprint image enrolled by the user, and a correlator that compares a live scan against the stored reference and signals a match. On a match, the fob generates a numerical value (a cyclic redundancy code derived from the stored reference image), encrypts it, and wirelessly transmits it to a "door" (car, room, building, ATM, computer). The user must "register" the value at each door, and the door compares the received value against the one stored during registration before granting access. The fob has both a verification mode switch and an enroll mode switch, and uses a stored private encryption key in a fixed public/private key pair.
• Why it matters: This is the single closest reference to the "key side" of the '954 claims. It discloses: portable device → stored biometric → local comparison → wireless transmission of a code derived from the biometric → external verification against a registered/authenticated value.
• Potentially anticipates: Claim 1 (biometric storage + local match + wireless code send), claim 12 (integrated device/validation + RF module), claim 22 (system), and arguably claim 5 (fingerprint) and claim 26 (biometric types).
• Gap: The "door" is not a third‑party‑operated trusted authority distinct from the application, and there is no separate secret decryption value stored with a device‑ID code. These are the likely differentiators Proxense would rely on.

2. US 6,219,439 B1 — "Biometric authentication system"

• Inventor: Paul M. Burger. Filing date: July 9, 1999 (prov. 60/092,196, July 9, 1998); Issued: April 17, 2001.
• Description (verified): A smart card storing biometric data (fingerprint, retina, voice, saliva) plus a non‑biometric identifier works with a portable reader that performs on‑board comparison of a live scan to the stored data. Crucially, the system "prevents communication with external sources prior to user authentication being confirmed," and performs a second comparison between the card's non‑biometric identifier and "other user data stored remote from the smart card" to further determine authentication status. The card chip stores the biometric data; the reader is self‑contained.
• Why it matters: Discloses a portable, tamper‑resistant token storing biometrics + an identifier, a local match, and a two‑stage authentication that includes a remote check — conceptually close to the '954's "device ID authenticated by a remote authority."
• Potentially anticipates: Claim 1, claim 12, claim 22; also relevant to claim 5 (fingerprint/retina/voice) and claim 26.
• Gap: Remote data is described as "stored remote from the smart card," not a third‑party‑operated trusted authority with a list of legitimate device IDs, and the application does not itself receive the access message.

3. US 4,993,068 A — "Unforgeable personal identification system"

• Assignee: Motorola, Inc. Issued: February 12, 1991.
• Description (per the '954's own citations and Burger's background): A portable memory device (credit‑card size) storing identification credentials such as facial photo, retina scan, voice and fingerprints, in encrypted form. The user presents the device and submits the physical characteristic(s) for comparison to determine whether to allow or deny access to a remote site.
• Why it matters: Foundational reference for "portable device storing biometric credentials + encrypted, presented for a remote allow/deny decision." Frequently cited in this art.
• Potentially anticipates: Broadly relevant to claim 1 and claim 22; weaker on the RF/wireless transmission and on the third‑party trusted authority.
• Gap: No described local RF transmission of a device‑ID to an authentication module/application as claimed.

Tier 2 — Key §103 references / secondary §102 candidates

4. US 6,363,485 B1 — "Multi‑factor biometric authenticating device and method"

• Assignee: Entrust Technologies Limited. Filing date: Sept. 9, 1998; Issued: March 26, 2002.
• Description (verified): A two‑factor biometric authenticator that generates error‑correction‑based partial encryption‑key seed data from a sampled biometric, stores that seed (not the secret key), and on later biometric input reconstructs a secret encryption key that is output to a host. Because it does not store the raw biometric or the secret key, the patent asserts the device "does not need to be tamper proof."
• Why it matters: Directly relevant to the '954 claims' "secret decryption value" and to a secret/key generated from biometric input — the specific element that distinguishes '954 from Hsu/Burger.
• Potentially anticipates/renders obvious: Element of claim 1 / 12 / 22 (secret decryption value). Likely a §103 reference in combination rather than standalone anticipation.
• Note: This reference teaches away from tamper‑proof storage ("does not need to be tamper proof"), which could be argued both ways.

5. US 6,016,476 A — "Portable information and transaction processing system and method utilizing biometric authorization and digital certificate security"

• Assignee: International Business Machines Corp. Issued: January 18, 2000.
• Description: A portable device combining biometric authorization with digital certificate security for information/transaction processing.
• Potentially anticipates: Broader §103 support for claim 1 (portable biometric‑authorized device) and claim 22 (system).

6. US 6,256,737 B1 — "System, method and computer program product for allowing access to enterprise resources using biometric devices"

• Assignee: Bionetrix Systems Corporation. Issued: July 3, 2001.
• Description: Centralized system permitting access to enterprise resources using biometric devices — relevant to the "authentication of the device/identity by a remote/central authority" element.
• Potentially anticipates: Claim 16 ("reader/verifier side" requesting authentication of a received identifier against a registered authority) and claim 22.

7. US 6,983,882 B2 — "Personal biometric authentication and authorization device"

• Inventor: Cassone. Issued: January 10, 2006 (appears in the '954's cited references).
• Description: A personal biometric authentication and authorization device carried by a user, performing biometric verification and authorizing transactions.
• Caution: Its issue date (2006) is after the '954 priority (Dec. 2004); whether it qualifies as prior art depends on its filing date (which I did not verify). Treat as a §103 reference only if it predates Dec. 20, 2004 — flagged as unverified.

Tier 3 — Contextual / secondary references (from the cited‑by and citation lists)

These appear in the '954 citation record and are relevant to specific dependent claims or to the gaming/age‑verification embodiments:

IPR‑cited prior art (invalidation‑track references)

From the PTAB filings retrieved, invalidity has been run primarily through §103 obviousness combinations, not §102 anticipation:

• IPR2024‑00232 (against sibling '730 patent): Grounds based on Ludtke, and Ludtke in view of Kon, with a stipulated dispatch of "Ludtke and Kon" as the core references. The Board's decision I retrieved discusses the "third party trusted authority" and "receiving an access message allowing access to an application" limitations — the same dispute that carries into the '954 claims.
• Other Proxense IPRs: Petitions relied on Cheshire (primary), with Meenan and Spurgat as secondary references (these appear in a related Proxense proceeding).
• Flagged uncertainty: I was not able to retrieve the full bibliographic citations for "Ludtke," "Kon," "Cheshire," "Meenan," or "Spurgat" (publication numbers, dates, and assignees). I am therefore not mapping them to specific '954 claims or asserting §102 status. They are cited here only to show the shape of the invalidity case (obviousness combinations) rather than anticipation.

Summary mapping (best‑evidence view)

Bottom line

• The two strongest §102 candidates for the '954 independent claims (1, 12, 22) are US 6,041,410 (Hsu/TRW) and US 6,219,439 (Burger) — both disclose the core sequence of portable biometric storage → local match → wireless release of an identifier → external verification against a registered value. US 4,993,068 (Motorola) is the deepest‑rooted supporting reference.
• The elements most likely to survive as the '954's distinguishing features are (a) the code being authenticated by a third party that operates a trusted authority, (b) the application itself receiving the access message, and (c) the stored write‑once "secret decryption value" alongside the device‑ID — for which US 6,363,485 (Entrust) is the closest art, typically via §103.
• I recommend verification before reliance on: the filing date of US 6,983,882, and the full citations of the IPR references Ludtke, Kon, Cheshire, Meenan, and Spurgat, which I could not retrieve. I did not fabricate any of those identifiers or dates.

I'll confirm the content of the key prior-art references cited on the patent page before constructing the obviousness analysis.

Obviousness Analysis — US 8,886,954 B1 under 35 U.S.C. § 103

Analyst note on scope and sources. This analysis uses the prior-art references listed in the "Citations (158)" and "Cited By (16)" panels of the US 8,886,954 page (the record art), plus the reference content I verified via search. All references discussed predate the December 20, 2004 effective priority date and are therefore available as § 102/§ 103 prior art. This is a technical obviousness assessment, not a legal opinion on validity; the ultimate question is before the PTAB and the Federal Circuit (see cross-reference at the end).

Framing / level of ordinary skill (PHOSITA). As of late 2004, a PHOSITA in this field would have been a person with a bachelor's degree in electrical/computer engineering (or equivalent) and 2–4 years of experience in biometric authentication, RF/wireless access control, and cryptographic credential handling — or equivalent experience. That person is presumed familiar with portable "fob"-style biometric authenticators, wireless (RF/IR) identity confirmation, write-once/tamper-resistant memory, and network-based credential verification services.

Contradiction flag (per the cross-reference instruction). The prior generated "Enforcement/PTAB/CAFC" section reports a live dispute over the construction of the claim phrase "a third party that operates a trusted authority." That construction issue is central to the obviousness analysis below, because the third-party trusted-authority element is what distinguishes the independents from the closest art (the TRW fob). I treat the phrase literally and note where a broader or narrower construction changes the analysis.

1. The limitations that do the work

The four independent claims (1, 12, 16, 22) reduce to a common stack of elements:

Elements A–C are squarely in the pre-2004 art. The definable novelty, if any, is (i) the specific package of write-once storage of a device ID plus a secret value, and (ii) the "third-party trusted authority → access message to the application" architecture of elements D–E.

2. Key references (from the record) and what each teaches

• US 6,041,410 (Hsu et al., TRW; examiner-cited). The single most material reference. Discloses a portable fob containing a fingerprint sensor, storage means for reference biometric data, an image correlator that compares the sensed print to the stored reference, and, on a match, a transmitter that wirelessly sends "user identification data" together with an encrypted numerical value derived from the stored reference image. The fob also stores a private encryption key "secretly in the device," and uses it to doubly encrypt the value. Before first use the user "registers" at the door, which stores the user's identifier and CRC for later comparison. → Teaches A, C, and the "register a code with the verifier" idea, plus a secret key stored on the device (the "secret decryption value").

• US 4,993,068 (Piosenka et al., Motorola; cited by examiner). "Unforgeable personal identification system": generates encrypted (one-way) biometric credentials stored on a portable memory device (credit-card size); at a remote site the reader reads the credentials, the user inputs biometrics, and a comparison determines access; claims recite fingerprint, retinal, voice, video (face), and pressure-tablet signature biometrics, and RSA-type public-key encryption with decryption at the remote site. Privilege text data is stored alongside. → Teaches A (encrypted/tamper-resistant storage of biometric credentials on a portable device) and the enumerated biometrics of claims 5/26.

• US 6,016,476 (Maes & Sedivy, IBM). Portable PDA with a biometric sensor (finger/thumb/palm print, handwriting, retinal), storing financial/personal data, with wireless RF and IR communication; a central server of a service provider (a third party) verifies the user and issues a temporary digital certificate; a transaction terminal establishes a link to the central server to confirm that the user was properly verified and to receive an authorization/verification result. Explicitly extends to medical information, financial accounts, and to implementation in a cellular phone or laptop. → Teaches elements D and E (third-party authority; the terminal/application obtaining the verification result over a network) and the application/device-form lists of claims 6, 7–9, 19–21, 25, 27–29.

• US 6,363,485 (Adams, Entrust; examiner-cited). Multi-factor biometric authenticator. Its background discusses a tamper-proof box that stores a secret symmetric/encryption key and self-destructs on tampering, and the invention concerns storing biometric-derived keying material internally to the device. → Teaches B (tamper-proof storage of a secret on the biometric device) and element A's "secret decryption value."

• US 6,704,333 (Bradford et al., Sierra Design Group). Two-authenticator player identification where the second authenticator is biometric data, usable at a game device; explicitly addresses age verification and the "open environment" problem (confirming an underage person is not playing), and lists RFID tags and handheld devices with RF/IR interfaces as authenticators. → Teaches age-verification (claims 3, 17) and casino-machine applications (claims 7, 19, 27), and the open-loop architecture.

• US 2002/0073042 A1 (Maritzen). Method/apparatus for secure wireless interoperability and communication between access devices. → Teaches wireless exchange and secure-channel establishment (claims 10, 18).

• US 2002/0173848 A1 (Imprivata). "Biometric authentication for remote initiation of actions and services." → Teaches that a successful biometric verification can trigger a code that remotely initiates a service — reinforces element C/D linkage.

• US 6,219,439 (Burger) and US 6,253,6737 (Bionetrix) — biometric authentication systems with centralized/enterprise verification, reinforcing the third-party-verification element.

• US 2002/0140542 A1 (Prokoski), "Personal biometric key" — reinforces the "biometric key" framing.

3. Obviousness of the independent claims

Claim 1 (key-side method)

Proposed primary combination: US 6,041,410 in view of US 6,016,476, further in view of US 4,993,068 (and optionally US 6,363,485).

Element-by-element:

• A (biometric + device ID + secret value on device): '410 teaches a fob storing reference biometrics, transmitting user identification data (the device/identity code) and storing a private encryption key (the secret value). '068 teaches the portable device storing encrypted identification credentials including privilege data.
• B (write-once/tamper-proof): '068 teaches one-way-encrypted immutable biometric credentials on the portable medium; '485 teaches tamper-proof storage of a secret key on the biometric device. Making the storage element physically write-once (WORM/ROM) is an obvious design choice to prevent substitution of biometric templates — a well-known hardening step, and the '954 specification itself frames tamper-proofing as a reliability feature, not an inventive leap.
• C (local compare + wireless send): squarely '410 (correlator + transmitter).
• D (send for authentication to a third-party trusted authority): not in '410 (the fob and door verify each other locally). This is supplied by '476, where a service-provider central server (a third party) verifies the user and issues/validates credentials, and the terminal requests authentication from that server.
• E (application receives access message from the trusted authority): also '476 — the transaction terminal links to the central server, which returns the verification/authorization result, and the terminal then permits the transaction.

Motivation to combine (KSR): '410 itself states the goal of reliable, convenient identification and warns that "loss [of the fob] may render the vehicle just as vulnerable to theft." A PHOSITA seeking the recognized (and documented) goal of open-loop authentication — where the relying party does not already share a pre-registered secret with the user — would naturally move the trust decision from the local "door" (as in '410) to a networked third-party verifier — precisely the architecture '476 already supplies (central server issuing/validating credentials, terminal requesting confirmation). Both references are in the same field of biometric access/transaction security, address the same problem (verifying the person, not just the token), and their combination is the predictable arrangement of known elements with no change in function. The age/market pressure is documented in '333 (casinos need to verify age in an open environment) and '476 (financial/medical transactions).

Secondary consideration that cuts toward obviousness: the write-once storage element is a matter of design choice; using ROM/one-time-programmable memory to protect biometric templates was conventional (see '068's immutable encrypted credentials).

Claim 12 (integrated device)

Same combination, mapped to apparatus:

• persistent storage media holding biometrics + device ID + secret value → '410 (memory + private key) and '068 (portable encrypted credential store);
• validation module receiving scan data and comparing → '410 correlator;
• RF communication module receiving an access message from the trusted authority → '476 (RF/IR capable PDA; terminal receives central-server confirmation). US 2002/0073042 (Maritzen) supplies explicit teaching of wireless interoperability between access devices for the RF module.

Claim 15 (LED prompt) is a routine user-interface choice; '410's user interface/indicator and '333's single activation button show prompting the user is conventional.

Claim 16 (verifier/reader-side method)

Proposed combination: US 6,016,476 in view of US 6,041,410, further in view of US 6,704,333 (optionally US 2002/0073042 for the secure wireless channel).

• "Wirelessly receive codes including a device ID from a biometrically verified user's device, the device ID registered with a third-party-operated trusted authority" → '410 (wireless receipt of ID + biometric-derived value that was registered at the relying device) + '476 (registration/verification at a third-party server).
• "Request authentication of the integrated device by the trusted authority; receive, at an application, an access message" → '476 (terminal requests confirmation and receives the result).
• "Allow the user access" → '476/'410.
• Age/DOB registration (claim 17) → '333 (age verification in gaming) and '476 (DOB stored at the server during enrollment).

Claim 22 (system)

Proposed combination: US 4,993,068 in view of US 6,016,476, further in view of US 6,041,410. The "integrated hardware device" + "authentication circuit" architecture is the straightforward combination of '410's portable biometric device with '476's server-verifying terminal. Claim 23 (device requests a scan in response to an authentication request from the circuit) is taught by '410's "wake-up" message from the door that initiates operation of the fob; claim 24 (no code if no verification) is inherent in '410's "no-match → no further processing."

4. Dependent claims

None of these dependents adds an element that is not either expressly disclosed or a predictable design choice in the record art.

5. Why a PHOSITA would have combined these references (motivation)

1. Same field, same problem. '410, '068, '476, '485, '333, and '042 all address verifying a person (not merely a token) before granting access to protected property, transactions, or entitlements. Combination of references within the same field to solve the same problem is the paradigm of obviousness.

2. Art-recognized deficiency in local-verification designs. '410's local fob-to-door scheme requires pre-registration at each door and shares secrets locally. The record art ('476, '439, '6737) had already moved trust to a central/third-party verifier to enable use across many relying parties — the "open loop" need that '333 independently identifies for casinos. The '954 patent's own stated advantage ("the integrated device can be authenticated for an application that is open to the public (i.e., in an open loop system)") is precisely the problem '476/'333 already solve, which undercuts any argument that the combination is non-obvious.

3. Predictable combination with expected results. Substituting a networked verifier for the local "door" verifier in '410, and hardening the on-device store with write-once/tamper-proof memory (known from '068/'485), involves no change in the respective functions of the elements and yields nothing more than the expected benefit of stronger, more flexible authentication.

4. Market/design pressure. Documented demand for eliminating memorized credentials and preventing use of stolen keys ('410 background; '476 background; '954 background) supplies the "reason to combine" that KSR requires.

6. Counterarguments a patent owner could raise (and my assessment)

• Teaching away / "third party" requirement. '410 and '068 contemplate the verifier (door/remote site) performing authentication, not an independent third-party authority. Proxense could argue the references teach away from inserting a third party and that "a third party that operates a trusted authority" is a specific architectural limitation. Assessment: '476 directly teaches a third-party service provider's server as the trust root, so this is met head-on; the dispute is really one of claim construction, which is where the PTAB/CAFC fight sits (see §7). A narrow construction of "third party" (an entity distinct from both device and application) is still met by '476's service-provider server.

• Write-once ("not capable of being subsequently altered") vs. encryption. '068 stores encrypted credentials but does not expressly say the medium is one-time-programmable. Proxense could argue this is a structural limitation absent from the art. Assessment: weak — '068's one-way encryption plus '485's tamper-proof key storage, combined with the common knowledge that ROM/WORM memory prevents template substitution, makes this an obvious design choice. Still, this is the element most worth pressing if one wanted to argue non-obviousness.

• Unexpected results / secondary considerations. I found no evidence of unexpected results or industry praise in the record text of '954 itself; the specification describes the write-once storage as a reliability (anti-tamper) feature, which is a known benefit, not an unexpected one.

7. Caveats and uncertainty flags

• Reference-content reliance. My characterizations of '410, '068, '476, '485, and '333 come from the verified abstracts/claim sets quoted above. Specific column/line cites should be confirmed against the full PDFs before any filing.
• The controlling live issue is claim construction, not art coverage. As flagged in the prior section, the PTAB (per the ip fray report carried on the patent page) addressed the construction of "a third party that operates a trusted authority," and Proxense has appealed (CAFC dockets 25-2168 and 26-1283, docket numbers taken literally from the page's litigation listing). The entire "D–E" half of my analysis hinges on how that phrase is construed; a construction that reads "third party" narrowly could still be satisfied by '476, but a construction that imports additional requirements (e.g., a dedicated standalone authority separate from the transaction network) could weaken the '476-based showing. Treat the mapping of each CAFC docket to a specific IPR as unverified (consistent with the prior section's flag).
• "Trusted authority" pre-2004 art is abundant. Because '476 and '6737 (2000–2001) already teach third-party/central verification of biometric users, the third-party element is unlikely to save the claims on its own.
• I did not verify whether US 5,131,038 (portable authentication transceiver with encrypted parametric memory) is formally in '954's citation list; I therefore did not rely on it, though if it is of record it would further strengthen the "portable device + encrypted biometric-in-memory + wireless interrogation" showing.

Bottom line: On the record art, the strongest § 103 combination against the independent claims is US 6,041,410 (TRW fob) as the primary reference, in view of US 6,016,476 (IBM/Maes central-server biometric authorization) and US 4,993,068 (Motorola unforgeable portable biometric credentials), optionally with US 6,363,485 (Entrust tamper-proof key storage) and US 6,704,333 (Sierra age/identity verification) for the dependents. Each element of claims 1, 12, 16, and 22 appears expressly taught or is a predictable design choice, and the motivation to combine is supplied directly by the art's own statements of purpose (open-loop authentication, stolen-key prevention, age verification). The principal open variable is the construction of "a third party that operates a trusted authority," which is being litigated on appeal and which a broader reading of '476 would satisfy.