US 10046696

I'll search for the specific patent number and related litigation records.

Let me search for litigation and CAFC dockets specifically involving this patent.

Let me retrieve the actual claim text for this patent.

US Patent 10,046,696 — Analyst Summary

Important scope note: I searched for the literal number 10046696 / US10046696B2. Search coverage is imperfect — I could not directly query the CAFC docket system, PACER, or USPTO PatentCenter/Assignment APIs, and the Google Patents page for this patent (my authoritative source, fetched 2026-09-29 from https://patents.google.com/patent/[US10046696](/patent/US10046696)/en) was truncated before the verbatim claims section. Where I lack authoritative text, I flag it below.

---

1. Bibliographic data (from the authoritative Google Patents record)

Continuity chain (per the patent's own "Related Application Data"): 15/348,766 is a continuation of 15/009,383 (filed 2016-01-28) → continuation of 13/770,970 (filed 2013-02-19, issued as US 9,248,777) → continuation of 13/191,404 (filed 2011-07-26, US 8,378,805) → continuation of 12/360,081 (filed 2009-01-26, US 7,986,223) → continuation of 11/046,047 (filed 2005-01-28, US 7,482,916) → provisional 60/553,426 (filed 2004-03-15). Term therefore runs from the 2005 non-provisional, which is why the record shows expiration on/about 2025-01-28 — i.e., the patent was already expired when the 2026 suits were filed, which materially limits recoverable damages.

---

2. Abstract (verbatim)

"An apparatus for use in a car includes: an input configured to receive data from a sensor; a processor configured to perform a statistical analysis using the data to determine a control parameter for controlling an automatic control of the car; and an output for providing the control parameter for controlling the automatic control of the car."

---

3. Independent claims — plain language

⚠️ Caveat on confidence. The Google Patents "Definitions" block for this patent (which is auto-derived from the claims) recites two separate "An apparatus for use in a car" independent claim sets, both sharing the same dependent-claim ladder (image data from a camera; spatial relationship to the lane; receiving vehicle speed; wireless network information; flat-panel display; control signal from a lever). That is consistent with the two-column table in the patent's own Summary section of the specification, which likewise states two independent "An apparatus for use in a car…" paragraphs. I do not have the verbatim, numbered claim text, so claim numbering is my inference, not a quoted fact.

Independent claim 1-type (broad):
A vehicle apparatus with (a) an input receiving sensor data, (b) a processor that performs a statistical analysis on that data to derive a control parameter for an automatic vehicle control, and (c) an output delivering that control parameter to the automatic control. Dependents narrow this to camera image data, lane-position/spatial-relation processing, speed input, wireless-network instructions, flat-panel display output, and a lever control signal.

Independent claim (narrower): Same structure, but the analysis determines a control parameter for an automatic control that involves a turn signal of the car.

Discrepancy to note: A 2026 district-court complaint analysis (Ex Parte case report for N.D. Tex. 3:26-cv-01372) characterizes claim 1 as also requiring "wherein the processor has machine learning capability," and paraphrases the control target as "a function of the car" rather than "an automatic control." The patent's abstract and the Google Patents Definitions block contain no express "machine learning capability" limitation in claim 1 and use "automatic control." The machine-learning/adaptive subject matter is clearly in the specification (FIG. 7 learning-capability system; processor 664 storing sway distance 802 and doing statistical/histogram analysis to adjust thresholds 510/512). I flag this as unresolved — either the complaint analysis paraphrased loosely, or claim 1 contains language not reflected in the Google Patents Definitions extract. Do not rely on the "machine learning capability" limitation without pulling the granted claims from PatentCenter.

---

4. Litigation findings (2026)

District court (asserted, alongside a family of related patents): Wyoming Technology Licensing (identified by Unified Patents as an NPE/entity of Leigh M. Rothschild; Rule 7 corporate parent disclosed as Patent Asset Management, LLC) asserted the '696 patent together with '916, '223, '805, '777, '343, '743, and '700 in a coordinated April 2026 campaign.

• W.D. Tex. 1:26-cv-00810 — Wyoming Technology Licensing, LLC v. General Motors LLC, filed 2026-04-01; Exhibit F = '696 Patent, Exhibit N = '696 Claim Chart. Stayed for settlement 2026-06; voluntary dismissal 2026-07-16; case closed 2026-07-17. (https://dockets.justia.com/docket/texas/txwdce/1:2026cv00810/1172898954)
• E.D. Tex. 2:26-cv-00268 — v. Toyota Motor North America, Inc., filed 2026-04-01, Judge Gilstrap. (https://litigation.rpxcorp.com/litigation/txedce-[244776](/patent/244776)-wyoming-technology-licensing-v-toyota-motor-north-america)
• N.D. Tex. 3:26-cv-01372 — Wyoming Technology Licensing, Inc. v. Volkswagen Group of America Inc, filed 2026-04-28; accused product is VW IQ.DRIVE / Travel Assist assisted lane changing; resolution/stay filed 2026-06-19. (https://dockets.justia.com/docket/texas/txndce/3:2026cv01372/[418895](/patent/418895); complaint analysis: https://ai-lab.exparte.com/case/dct/txnd/3:26-cv-01372/doc/analysis/1)
• The Google Patents record also lists N.D. Tex. 3:26-cv-01371, N.D. Tex. 4:26-cv-00451, and the W.D. Tex. and E.D. Tex. cases above, via Unified Patents litigation data (https://portal.unifiedpatents.com/litigation/...), plus a Darts-ip "first worldwide family litigation" link (family 34921984).
• Related: Unified Patents published prior art against claim 33 of US 7,482,916 — the '696's earliest non-provisional ancestor — on 2026-05-19 (https://www.unifiedpatents.com/insights/2026/5/19/wyoming-technology-vehicle-patent-prior-art-found).

CAFC 2026 docket: no results found. My searches for a Federal Circuit 2026 appeal involving 10046696 returned nothing relevant. I cannot confirm this negative with high confidence, because I was unable to query the CAFC docket (CAFC-BBS/appellate CM-ECF) directly and one targeted query returned zero hits. Do not treat "no CAFC appeal" as established — verify against the Federal Circuit docket if that point matters. As of the records I could retrieve, the '696 dispute activity is at the district-court stage only.

---

5. Bottom line

US 10046696 is a late-chain continuation (filed 2016-11-10, issued 2018-08-14) of Anita Au and Gerald Chan's long turn-signal automation family reaching back to a 2004 provisional. Its claims are apparatus claims to sensor-data-in → statistical-analysis-in-processor → control-parameter-out for automatic vehicle control, with a narrower independent claim tying that control to the turn signal. It is now owned by Wyoming Technology Licensing LLC (ex-AutoSignal LLC, ex-individual inventors), was asserted against GM, Toyota, VW and others in the 2026 Texas campaign, and is expired as of 2025-01-28. Two specifics I could not resolve authoritatively: (a) verbatim numbered claim language — particularly whether claim 1 contains a "machine learning capability" limitation, and (b) whether any Federal Circuit 2026 appeal exists.

I'll search litigation databases for this specific patent number.

Let me search for the remaining Texas cases and any additional defendants.

Let me check the Unified Patents caselist and look for any additional defendants or earlier suits.

Known litigation involving US Patent 10,046,696

Scope/date note first — a contradiction to flag: Your task header states "Current Date: April 26, 2026," but the authoritative Google Patents record, the search results themselves, and the prior analysis session are all dated later (the patent page was fetched 2026-09-29, and docket entries below run through late September 2026). I am therefore reporting the docket as it stands through approximately 2026-09-29, and I flag that the "April 26, 2026" date in the task header cannot be reconciled with the retrieved records. I have not auto-corrected any case number, party name, or patent number.

Important statutory context that applies to every case below: the '696 patent is a late continuation of the Au/Chan turn-signal family whose term ran from the 2005-01-28 non-provisional, so it shows as expired as of 2025-01-28. All 2026 suits are therefore necessarily pre-expiration damages suits (the Hyundai complaint reportedly says exactly this — damages "through the patent's expiration on January 28, 2025"). (Korean press summary)

---

Table of identified cases

---

Family / campaign context

Unified Patents identifies Wyoming Technology Licensing as an NPE and entity of Leigh M. Rothschild, and states the related '916 patent (the '696's ancestor) "has been asserted against General Motors, Toyota, Mercedes-Benz, Honda, and Volkswagen" — consistent with cases 1–5 above. (Unified Patents, 2026-05-19) The GM complaint enumerated the whole family in one action: '916, '223, '805, '777, '343, '696, '743, '700. Unified Patents also published prior art against claim 33 of US 7,482,916 on 2026-05-19 (same link).

What I could NOT establish (please verify before relying on it)

1. Federal Circuit: No 2026 CAFC appeal involving 10046696 appeared in any result. I could not query the CAFC docket directly, so this is an unverified negative, not a confirmed absence of appeals.
2. Pre-2026 suits: An aggregate party page shows a Wyoming Technology Licensing case(s) with a stipulated dismissal with prejudice on 2024-07-31 and an earlier dismissal order of 2024-07-13 (Texas, Judges Pitman/Payne), but I could not obtain the case number, court, defendant, or whether the '696 was asserted. Flagging as an unresolved lead. (Ex Parte party page)
3. Case 6 (Hyundai) case number and case 7 (BMW) district/'696 involvement were not captured before my search budget ran out.

Corroboration of the previously flagged claim-language discrepancy

Your earlier section flagged that a complaint analysis paraphrased claim 1 of the '696 as requiring "wherein the processor has machine learning capability," which the Google Patents "Definitions" extract did not reflect. That flag is now corroborated by two independent complaint analyses:

• VW (3:26-cv-01372): "Claim 1 of the '696 Patent recites… An apparatus for use in a car comprising an input configured to receive data from a sensor. A processor configured to perform a statistical analysis using the data to determine a control parameter for controlling a function of the car. An output for providing the control parameter for controlling the function of the car. Wherein the processor has machine learning capability."
• Toyota (2:26-cv-00268): "…performs a 'statistical analysis' using the data to determine a control parameter for an automatic vehicle function. The claims also recite machine learning capability."

Both complaints thus treat the machine-learning limitation and "function of the car" phrasing as claim-1 text, differing from the abstract's "automatic control of the car." This does not resolve the issue authoritatively (complaint paraphrase ≠ granted claim text), and no verbatim granted claim set has yet been pulled from PatentCenter in this analysis. Do not plead or rely on either formulation until the granted claims are confirmed.

Bottom line: Eight identified '696-related matters, seven of them the 2026 Rothschild/Wyoming Texas–California campaign (GM [closed], Toyota [open], Honda [open], VW [settled/dismissed], Mercedes-Benz [open, trial 10/2027], Hyundai [new, Aug 2026], BMW [unconfirmed]), plus an unverified 2024 dismissal lead. Plaintiff-side counsel appearing throughout is Rene A. Vazquez (with Joseph J. Zito); the campaign is pre-expiration damages only, since the patent lapsed 2025-01-28.

I'll verify the structured "no PTAB activity" finding against public sources before writing.

Proceedings overview

Total AIA trial proceedings on file for US 10,046,696: 0 (zero). The canonical structured source for this task — the USPTO Open Data Portal PTAB block — reports no AIA trial proceedings (no IPR, no PGR, no CBM) for this patent as of the most recent ingest, and my independent web searches for "10046696" + IPR/PTAB returned no proceeding number of any kind. The breakdown is therefore trivially: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. No ex parte reexamination request surfaced either.

Defensive posture for a defendant today: this is the "claims are untested" scenario, not the "patent has survived two IPRs and is hardened" scenario — but with an important twist. The '696 is expired (expiration recorded 2025-01-28, consistent with the family's term running from the 2005-01-28 non-provisional), so a defendant is not facing injunctive relief; the exposure is past damages only. That cuts both ways: it lowers the urgency of an IPR, but it also means there are no § 315(e)(2) estoppel straps on any prior art — every ground you want to raise is still available in district court. The absence of any PTAB filing is itself notable given that the patent has been asserted against GM, Toyota, Mercedes-Benz, Honda, Volkswagen, Hyundai, and Lucid in the 2026 campaign (compare the '696's own ancestor, US 7,482,916, against which Unified Patents published an invalidity chart on 2026-05-19 — see below). Nobody has pulled the AIA trigger yet.

---

(No proceeding) — no AIA trial exists on US 10,046,696

There is no IPR20XX-XXXXX, PGR20XX-XXXXX, or CBM20XX-XXXXX to report. Rather than invent one, here is what the record actually shows, plus the nearest-adjacent activity:

• Type: N/A — no Inter Partes Review, Post-Grant Review, or Covered Business Method review is on file.
• Filed: N/A.
• Status: N/A ("no AIA trial proceedings" per the USPTO ODP block). Plain-English gloss: the Board has never been asked to review any claim of this patent.
• Judge panel: N/A.
• Petition grounds: N/A. No petitioner has challenged any claim under § 102, § 103, or § 112.
• Institution decision: N/A.
• Final Written Decision: N/A. No claim of US 10,046,696 has ever been canceled, confirmed, or construed by the PTAB.
• Settlement / termination: N/A.
• Appeal: No Federal Circuit appeal involving 10046696 was found. I could not query the CAFC docket (appellate CM/ECF) directly, so treat this as an unverified negative, not a confirmed absence. My searches surfaced nothing on CourtListener or in Federal Circuit opinion repositories. (Source format for checking: CourtListener, PTAB E2E, PTAB Decisions.)
• Defensive value: Nothing is foreclosed. With zero FWDs, § 315(e)(2) estoppel is a blank slate, and — because the patent is expired and the only live defense interest is past damages — the scope of art you can bring is limited mainly by the printed-publication requirement in an IPR, not by anything the Board has already decided.

Why the PTAB vehicles are narrow here (useful so you don't chase a dead end):

• PGR: unavailable. The PGR window is 9 months from grant; the '696 issued 2018-08-14, so the window closed on/about 2019-05-14. Separately, PGR reaches only patents with a claim having an effective filing date on/after 2013-03-16, and this family claims priority to 2004-03-15 — so PGR would very likely be unavailable on that ground as well.
• CBM: unavailable. The AIA transitional CBM program sunset 2020-09-16 for new petitions.
• IPR: the only available AIA vehicle. IPR of an expired patent is permitted (the Federal Circuit has allowed IPR to proceed against expired patents on the rationale that the claims remain enforceable for past infringement — Sony Corp. v. Iancu, 924 F.3d 1235 (Fed. Cir. 2019); verify the citation before relying on it, but the principle is that "the patent is expired" is not a bar to institution). Practical leverage is reduced because cancellation of an expired claim mainly buys you non-liability for past damages and a clean hands position, not injunctive relief.

Nearest-adjacent defensive activity in the family (not a PTAB proceeding): Unified Patents published on 2026-05-19 that it "successfully identified and charted prior art against claim 33 of U.S. Patent 7,482,916," the '696's earliest non-provisional ancestor, describing Wyoming Technology Licensing as an NPE and entity of Leigh M. Rothschild. The chart itself is public (invalidity chart PDF) and relies in part on US 2004/0143381 A1 ("'381") and a lane-change-aid system reference. This is prior-art scouting, not a filed petition — do not cite it as a proceeding, and do not assume claim 33 of the '916 maps onto any claim of the '696.

Cross-reference update from this search (litigation, not PTAB): a case not captured in the earlier litigation section surfaced — Wyoming Technology Licensing, LLC v. Lucid Group, Inc., E.D. Tex. 2:26-cv-00505 (Judge Gilstrap; Mag. J. Payne), where the complaint lists Exhibit B — '696 Patent and Exhibit E — '696 Patent Chart among three asserted patents ('343, '696, '916). PacerMonitor. Treat as a new but consistent data point in the same campaign.

---

Strategic summary

Claim status: entirely UNTESTED. No claim of US 10,046,696 has been canceled, confirmed, or even construed by the Board. Nothing is "dead," nothing is "hardened," and there is no PTAB claim-construction ruling to borrow. The patent's relevance to the 2026 campaign is as an expired apparatus patent asserted for past damages, and (per the earlier sections) with a live, unresolved discrepancy about whether claim 1 actually recites a "machine learning capability" limitation and whether the control target is phrased as "an automatic control of the car" (abstract/Google Patents Definitions) or "a function of the car" (VW and Toyota complaint analyses). That ambiguity is currently the single highest-value thing to resolve — it changes both the § 112 posture and the obviousness story — and it can be resolved by pulling the granted claims from PatentCenter rather than by any PTAB filing.

Estoppel landscape: clean. With no IPR, PGR, or CBM ever instituted, there is no § 315(e)(2) estoppel against anyone. No petitioner is barred from raising any ground in district court, and no petitioner's privies are contaminated. For a defendant now being asserted against (e.g., Toyota, Honda, Mercedes-Benz, Hyundai, Lucid), that means the full prior-art universe remains available — including printed publications, patents, and system/on-sale art — subject only to the ordinary § 102/§ 103 and § 282 burdens. The one clock to watch is § 315(b): a petition must be filed within one year of service of a complaint alleging infringement. For a party served on the ~2026-04-01/2026-04-28 wave, that bar runs out around 2027-04-01/2027-04-28; for Hyundai (~2026-08-24 per earlier reporting), around 2027-08-24. Note Thryv, Inc. v. Click-to-Call Techs., LP, 140 S. Ct. 1367 (2020): GM's voluntary dismissal (2026-07-16) does not un-ring the § 315(b) clock that started when GM was served — though GM's notice of settlement suggests the parties resolved rather than litigated.

Pattern signals. (i) No repeat petitioner — there is no petitioner at all. (ii) No patent-owner PTAB appeals — the owner has never had an FWD to appeal, and no CAFC activity was found (unverified negative). (iii) A defensive aggregator is in the family's chain but has not filed: Unified Patents published an invalidity chart against the ancestor '916's claim 33 on 2026-05-19 and has publicly targeted the Rothschild/Wyoming campaign, yet no IPR appears on file for the '696. (iv) The owner's pattern in district court is file-and-settle/stay: GM (stay 2026-06-19, voluntary dismissal 2026-07-16, closed 2026-07-17), VW (unopposed stay + resolution 2026-06-19, dismissed without prejudice 2026-06-22 with a 30-day reopening right), versus genuinely contested schedules (Mercedes-Benz discovery to 2027-05-21, trial 2027-10-18). This is a licensing campaign, which is consistent with the absence of any party having invested in an IPR.

---

Recommended next steps

If you are a defendant and want to know where you stand, in priority order:

1. Do NOT represent to a court or counterparty that any claim of the '696 has been invalidated. It has not. No FWD exists. There is no PTAB paper to quote, and I will not invent claim numbers or a disposition.
2. Resolve the claim-text question first. Pull the granted claims and full prosecution history for US 10,046,696 from USPTO PatentCenter (application 15/348,766). Confirm verbatim (a) whether claim 1 recites "wherein the processor has machine learning capability," (b) whether it says "automatic control of the car" or "function of the car," and (c) the exact claim numbering. Both the VW (3:26-cv-01372) and Toyota (2:26-cv-00268) complaint analyses assert the machine-learning limitation; the abstract and Google Patents Definitions extract do not. This is the gating issue for every invalidity and § 112 theory you would build.
3. Preserve your § 315(b) window. If you have been served, calendar the one-year bar (service date + 1 year). Even with the patent expired, an IPR remains the only AIA vehicle and can still be worth filing for a past-damages defense and for a clean § 315(e)(2) estoppel position against follow-on theories — but decide before the bar, because there is no pending petition you could join (no joinder path under § 315(c) exists without a first petition).
4. Copy the Unified Patents work product rather than re-inventing it. Review the public '916 claim-33 invalidity chart and its references (notably US 2004/0143381 A1), then check whether those references are within the '696's 2004 priority window and whether they read on the '696's claims specifically. Do not assume chart-to-chart transfer.
5. If no PTAB activity is your whole answer, say so plainly — and use the silence. The absence of any IPR is a meaningful signal on its own: this is a well-asserted family by a repeat NPE, and neither an equipment manufacturer nor a defensive aggregator has yet committed to a Board challenge. Combined with the 2025-01-28 expiration, your practical defenses are (a) the merits/validity fight in district court with a fully open art record, (b) § 101/§ 112 attacks that were never tested at the Board, and (c) damages ceilings tied to expiration — including marking and pre-suit-damages defenses under § 287 for any licensee/sublicensee chain.

Two open items to verify before relying on anything above:

• CAFC: no 2026 Federal Circuit appeal involving 10046696 was found, but I could not query the CAFC docket directly — this is an unverified negative. Check CourtListener and the Federal Circuit's own docket.
• Pre-2026 PTAB/reexam history: my searches covered IPR/PGR/CBM only. I did not exhaustively check for an ex parte reexamination (which would not appear in the AIA-trial ODP block). Check the '696's file wrapper in PatentCenter for reexam certificates before stating categorically that the claims have never been reviewed.

Date note (resolving an earlier flag): the prior session flagged a contradiction between a "current date of April 26, 2026" in its task header and records running through September 2026. This task header states 2026-09-29, which is consistent with the authoritative records (Google Patents fetch 2026-09-29; GM order closing case 2026-07-17; Mercedes-Benz scheduling order 2026-06-01). The earlier discrepancy appears to have been a stale header, not a data conflict.

I'll research the assignment record for US 10,046,696, starting with the USPTO Assignment Center and cross-referencing the sources I have.

Those weren't productive. Let me try more targeted searches for the assignment records and the correspondent attorneys.

Let me search for the specific assignment records and the correspondent attorneys on this chain.

I hit my search-step limit before I could pull the primary Assignment Center records directly. I'll report exactly what I could and could not verify, and I will not fabricate reel/frame numbers.

Retrieval limitation (read first): The USPTO Assignment Center (assignmentcenter.uspto.gov) and assignment.uspto.gov are JavaScript applications that my searches could not query directly, and no indexed mirror I reached exposes the reel/frame, execution dates, or correspondent for US 10,046,696's specific records. The only authoritative assignment data I have is the Google Patents "legal events" block for US10046696B2 (fetched 2026-09-29), which lists the two recorded conveyances but not their reel/frame numbers or correspondents. Everything below is marked accordingly. Do not treat any reel/frame as established — I cite none.

---

Inventors

Unusual pattern — flag: Both inventors apparently kept personal ownership of the family for ~11 years. The earliest non-provisional (11/046,047, filed 2005-01-28) issued as US 7,482,916 in 2009 with no company assignee of record, and it was not until 2020-03-12 that Au and Chan assigned to AutoSignal LLC. So the "all inventors depart the assignee within 12 months" fire-sale signal does not apply here in its usual form; instead this is the mirror image — independent inventors holding the patent personally through multiple continuations, then forming an LLC to hold it. I could not determine either inventor's employer at the time of filing; the patent gives no assignment to an employer and no employment data surfaced.

---

Original assignee

• Entity named on the issued patent: "Individual" (i.e., no corporate assignee at grant). Google Patents records the original assignee literally as Individual.
• Primary line of business / product: Because the original owner was the inventors personally, there is no operating-company assignee to assess at grant. The later-formed AutoSignal LLC (which took the assignment on 2020-03-12) appears to be the inventors' own holding/licensing vehicle rather than a manufacturing company — but I could not verify any AutoSignal product in commerce, and I could not confirm AutoSignal's state of organization, registered agent, or current status. Current status of AutoSignal LLC: unverified (I did not retrieve a dissolution, merger, or bankruptcy record).
• Product-embodying evidence: None found. The patents cover automatic turn-signal actuation from camera/lane data; I found no evidence that AutoSignal (or the inventors) ever shipped a commercial product embodying the claims.

---

Assignment timeline

Chronological, from the authoritative Google Patents legal-events record. Reel/Frame and Correspondent are NOT available in my retrieved data — they are shown as "not retrieved," not as blanks I filled.

• Executed [not retrieved] / recorded 2020-03-12 — Reel not retrieved/not retrieved

  • Conveyance: Assignment ("ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: Au, Anita; Chan, Gerald (the two inventors)
  • Assignee: AutoSignal LLC
  • Correspondent: not retrieved — could not confirm; flag if it recurs.
  • Context: Inventors transfer the family (including this '696) into their own LLC — a transfer-to-holding-entity, not an operating-company assignment.

• Executed [not retrieved] / recorded 2026-01-28 — Reel not retrieved/not retrieved

  • Conveyance: Assignment ("ASSIGNMENT OF ASSIGNOR'S INTEREST")
  • Assignor: AutoSignal, LLC
  • Assignee: Wyoming Technology Licensing, LLC
  • Correspondent: not retrieved — this is the single highest-value field to capture (see Signal 3).
  • Context: Transfer to an asserter — AutoSignal's family goes to a Leigh M. Rothschild NPE entity, recorded ~63 days before the first 2026 infringement suits.

No other post-issuance assignments appear on the '696 in the Google Patents legal-events block. In particular I see no recorded security agreement, license, merger, name change, or release for this patent. If Assignment Center shows additional entries (e.g., a pre-2018 inventor assignment, or a confirmatory/corrective filing), they were not surfaced to me.

Naming discrepancy to flag (carried forward, not resolved): the earlier litigation section noted the 2026 complaints use both "Wyoming Technology Licensing, LLC" (W.D. Tex., E.D. Tex.) and "Wyoming Technology Licensing, Inc." (N.D. Tex. cases). The Google Patents assignment record uses "WYOMING TECHNOLOGY LICENSING, LLC." I have not determined whether these are (a) the same entity styled inconsistently in pleadings, or (b) two distinct entities. Verify the entity form on the recorded assignment.

---

Timeline diagram

timeline
    title Ownership of US 10046696
    2004 : Provisional 60553426 filed
    2005 : Nonprovisional filed by Au and Chan
    2018 : Patent 10046696 issued
    2020 : Assigned to AutoSignal LLC
    2026 : Assigned to Wyoming Technology Licensing LLC
         : First 2026 infringement suits filed

---

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT.
Documented by the 2026-01-28 recorded assignment (Assignor AutoSignal, LLC → Assignee Wyoming Technology Licensing, LLC). The assignee is a licensing-only LLC with no products in commerce, and it is identified by Unified Patents as an NPE and entity of Leigh M. Rothschild, with a corporate parent disclosed as Patent Asset Management, LLC. The name suffix "Licensing" and the licensing-only function are backed by the assertion campaign (see Signal 5), not merely by the name. Caveat: I could not independently confirm a registered-agent-service address or single-member Delaware/Texas registration from Assignment Center, so I am grounding this call on the assignment record + Unified Patents' identification, not on the name alone.

2. Known asserter in the chain — PRESENT.
The current assignee, Wyoming Technology Licensing, LLC, is surfaced by Unified Patents (2026-05-19 insight) as an NPE/entity of Leigh M. Rothschild, and Rothschild's entities are documented as high-frequency plaintiffs — the NPE Showcase piece (reproduced in a W.D. Wash. filing, Valve Corp. v. Rothschild, 2:23-cv-01016) attributes 1,249 patent infringement lawsuits to his entities, with fee awards for baseless suits. Note the assignee is a Rothschild entity even though it is not on the task's named list (Acacia, Marathon, IV, Wi-LAN, etc.) — the catch-all "surfaced by Unified Patents or RPX as a high-frequency plaintiff" applies. Prior link: AutoSignal LLC does not match any listed NPE. Reel/Frame for the 2026-01-28 assignment was not retrieved.

3. Repeat correspondent across the chain — UNCLEAR (not captured).
This is the one signal I most wanted and could not confirm: I did not retrieve the correspondent of record for either the 2020-03-12 or the 2026-01-28 assignment. So I cannot state that a single attorney runs both links, and I will not do so. As leads only (not findings for this patent), publicly filed Rothschild-chain assignments show correspondents such as Woodcock Washburn LLP (Robin L. Parmelee) on a 2012 Rothschild Trust Holdings confirmatory assignment, Casella & Hespos LLP (Anthony J. Casella) on Trust Licensing LLC filings, and Jon M. Isaacson on a Leigh M. Rothschild filing — none of these is confirmed for US 10,046,696. Please pull the correspondent names from Assignment Center; recurrence there would be a strong finding.

4. Cascading transfers — NOT PRESENT (for this patent).
Only two recorded conveyances, and they are ~6 years apart (recorded 2020-03-12 and 2026-01-28) — not consecutive transfers through chained LLCs within 24 months, and no shared-correspondent address could be shown (see Signal 3). The tight timing is in the litigation, not the title chain.

5. Pre-litigation transfer — PRESENT.
The 2026-01-28 transfer to Wyoming Technology Licensing precedes the first 2026 suits (GM and Toyota filed 2026-04-01; Mercedes-Benz 2026-04-10; VW/Honda 2026-04-28) by roughly 2–3 months — well inside the 6-month window. A single-purpose asserter was interposed immediately before a coordinated multi-defendant campaign, which is consistent with arranging the chain to establish a clean standing/ownership record for assertion.

6. Bankruptcy fire-sale — NOT PRESENT.
No bankruptcy proceeding involving Au, Chan, or AutoSignal LLC surfaced. (Rothschild himself appears in a Chapter 11 context on an unrelated 1990s Intracorp matter, but that is not an assignor in this chain and is irrelevant to the '696.)

7. Privateering — NOT PRESENT.
Privateering requires an operating company transferring to an NPE that sues on its behalf. Here the transferor, AutoSignal LLC, shows no product and no evidence of an operating business; this looks like an inventor-owned holding vehicle ceding to an asserter, not a product company weaponizing its own patents. (Marked not present, but note it is a close cousin of Signal 1.)

8. Defensive aggregator — NOT PRESENT.
The chain ends at Wyoming Technology Licensing, LLC, an active asserter. It does not terminate at RPX, AST, LOT, Unified Patents, or OIN. The patent is not neutralized. (Note the inverse context: Unified Patents has published prior art against the ancestor '916's claim 33 — that is adverse scouting, not ownership.)

---

Verdict

NPE — high confidence.

Two strong signals plus a third are concretely grounded in the record: (1) the recorded 2026-01-28 assignment of US 10,046,696 from AutoSignal, LLC to the licensing-only NPE Wyoming Technology Licensing, LLC — a transfer-to-asserter with no product; (2) the assignee is publicly identified by Unified Patents as an NPE/entity of Leigh M. Rothschild, whose entity stable is documented as one of the highest-volume litigation filers in the country; and (5) that transfer was recorded only ~63 days before the first 2026 infringement suits were filed, squarely inside the pre-litigation window. The chain is inventors → AutoSignal LLC (their own vehicle, 2020-03-12) → Wyoming Technology Licensing LLC (2026-01-28), terminating at an active asserter, so it is neither an operating-company assertion nor defensively neutralized. Caveat on confidence: the correspondent and reel/frame fields are not yet captured, and the LLC-vs-Inc entity discrepancy for the 2026 assignee is unresolved; neither changes the two-signal NPE call, but both should be filled in.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent 10046696 / application 15/348,766), and the indexed view at https://assignment.uspto.gov/patent/index.html.

To finish this chain, retrieve these four fields: (i) reel/frame for both recorded assignments; (ii) the correspondent on each (Signal 3 is currently unresolved); (iii) the recorded execution dates (vs. recording dates) for the 2020 and 2026 assignments; (iv) the exact legal name of the 2026 assignee (LLC vs. Inc), plus any interim entries Assignment Center shows that Google Patents' legal-events block did not surface.

I'll search for the cited references and prior-art record for this specific patent.

Let me pull the specific citation record and the references named in the '696 specification.

Prior-art analysis — US 10,046,696 B2 ("Automatic control systems for vehicles")

Date note (carry-forward): this task header says 2026-04-26; the authoritative record and the prior sections in this analysis run through 2026-09-29. That discrepancy was already diagnosed earlier as a stale header. I flag it and do not auto-correct it.

Retrieval limitation that governs this whole section. The authoritative full text I hold for the '696 is truncated at the point where the pre-grant publication text ends, before the grant front page ("References Cited") and before the verbatim claim set. So I can report (a) the references cited in the specification body, which the supplied full text reproduces verbatim, and (b) references surfaced by searching elsewhere — but not the examiner-cited "References Cited" list on the '696's own front page.** That list is the primary prior-art record and I could not retrieve it. Anything below drawn from (b) is marked as provenance-limited.

---

A. Identifier discipline (per your "do not auto-correct" instruction)

Searching the literal string 10046696 pulls in two near-miss documents that are not this patent and must not be merged into this analysis:

I make no correction to either identifier.

---

B. References cited in the '696 specification (verbatim from the supplied full text)

The '696's specification names exactly six references, all in the sensor section, introduced as devices "for detecting a presence or an absence of a lane boundary" — i.e., cited as enabling-technology background, not as claim-1 anticipation targets.

I will not state titles, dates, or inventors for items 2–6 because I did not retrieve them. Note that five of the six are on their face lane-detection references, not turn-signal-actuation references — the '696 recites them as building blocks, which tends to make them background/§ 103 art rather than § 102 anticipation for the independent claims.

---

C. The strongest candidate prior art surfaced (provenance: Unified Patents' public chart against the ancestor)

Also surfaced, lower confidence: US 6,629,030 B2 (Klausner et al., issued 2003-09-30; family includes the colliding DE 10046696 A1) and US 2003/0168839 A1 (Mitoshi, 2003-09-11), both cited X/Y respectively in the WO 2007/017730 A3 search report. Neither was retrieved in substance, and neither is confirmed to concern turn-signal actuation. Their relevance is unverified; do not use them without pulling full text.

---

D. What I will affirmatively state about § 102

1. No reference I retrieved anticipates the '696's broad independent claim (input from sensor → statistical analysis in the processor → control parameter output for automatic vehicle control). The statistical/learning element — processor 664 storing the sway distance 802, building a distribution curve or histogram, and adjusting thresholds 510/512 (FIG. 7/FIG. 8) — is the feature the spec-cited lane-detection art plainly does not reach.
2. The turn-signal independent claim is the vulnerable one, and US 2004/0143381 A1 is the reference most likely to read on it under § 102 — subject to the date and claim-transfer verifications above.
3. Do not map claims by number yet. The earlier sections established an unresolved discrepancy: the abstract/Google Patents Definitions say "automatic control of the car," while the VW and Toyota complaint analyses quote claim 1 as "a function of the car" plus "wherein the processor has machine learning capability." Which formulation is the granted text changes every § 102 conclusion in this section — the "statistical analysis"/machine-learning limitation is the difference between easy anticipation and no anticipation. This gating issue is identical to the one flagged earlier and remains unresolved.

---

E. Verify these before relying on any of the above

1. Pull the '696 grant front page (PatentCenter, application 15/348,766) — the "References Cited" list, both U.S. and foreign, is the examiner's actual prior-art record for this patent and is completely absent from what I could retrieve. That is the single highest-value missing item.
2. Pull verbatim granted claims to fix the "automatic control" vs. "function of the car" and "machine learning capability" question.
3. Retrieve full text + exact dates for US 4,348,652; 5,979,581; 5,790,403; 5,957,983; 5,982,278; and US 2002/0175813, and for US 2004/0143381 A1 (publication date and U.S. filing date both matter).
4. Confirm the identity of the "US 10,469,966 B2" front page in the W.D. Tex. PDF before attributing any of its cited references to the '696.
5. Do not import the Unified Patents '916 claim-33 chart into a '696 analysis without re-mapping; it is adverse scouting against an ancestor, not a '696 invalidity contention.

I'll ground the obviousness combinations in verified reference data before building the analysis.

The '381's publication date (2004-07-22) is now confirmed and is pivotal. Let me verify its U.S. filing date and pull the remaining specification-cited references.

Key finding: the '381 was U.S.-filed 2003-11-05 (App. 10/700,675, DaimlerChrysler), with DE priority 2002-11-05 — that predates the '696's 2004-03-15 provisional. Let me pin down the last few references.

Obviousness Analysis — US 10,046,696 B2 under 35 U.S.C. § 103

Building on the prior-art section. This analysis uses the reference set that section assembled + the one reference it flagged as "the single most on-point § 102 candidate." I have now verified the critical bibliographic facts for that reference, and they change the strength of the analysis significantly. Per the "do not repeat" instruction, I do not re-cite the six specification-cited references in full; I cross-reference them.

Two carry-forward flags before the merits:

1. Date. This task header says 2026-04-26. My authoritative records and the prior sections in this analysis run through 2026-09-29. As previously diagnosed, the April header is stale; I use 2026-09-29.
2. Claim text is still unverified. The prior sections established an unresolved discrepancy — abstract/Google-Patents-Definitions say "automatic control of the car," while the VW and Toyota complaint analyses quote claim 1 as "a function of the car" plus "wherein the processor has machine learning capability." Every ground below is written to be robust to either reading, and I state where each formulation matters.

---

1. The reference-set foundation (verified this session)

The headline fact: the '381 is not merely close art on the turn-signal claim. Its U.S. filing date (2003-11-05) precedes the '696's earliest claimed date (2004-03-15), so unlike a 2004-07-22-publication-only reference, the '381 cannot be date-avoided by the '696's provisional priority. The prior section's caveat ("if the '381 published after 2004-03-15, it is not § 102(a) art") is thereby superseded: the '381 is available regardless.

---

2. Legal framework applied

• Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content; differences; PHOSITA level; secondary considerations).
• KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007): motivation may flow from the reference itself, the nature of the problem, the "design incentive," market demand, or common sense; a predictable variation of a known element using known methods is obvious; "obvious to try" can defeat patentability where the solution is one of a finite number of predictable options.
• Pre-AIA vs. AIA: the '696 issued 2018 from a 2016 continuation, but if every claim is supported by the 2005-01-28 non-provisional, pre-AIA §§ 102/103 govern (AIA § 3(n)(1)) and the effective filing date is 2004-03-15. If any claim limitation (e.g., a "machine learning capability" element) lacks written-description support in the 2004/2005 disclosure, AIA § 102/103 applies and the art universe expands enormously (2004–2016).
• Two-track posture: every ground below works under both regimes, because the '381 has both a 2002-11-05 effective date and a 2003-11-05 U.S. filing date.

PHOSITA (proposed): a person with a B.S. in EE/CS/ME plus ~2–3 years in automotive ADAS, or equivalent, familiar with camera/laser lane-marking recognition, Hough/template line-fitting, vehicle data buses (CAN), and turn-signal/switching circuitry as of 2003–2004. The '696's own specification sets this bar — it calls ASICs, Pentium-class processors, CCD cameras, Hough/curve-fitting, and stored templates "well known in the art."

---

3. Element-by-element mapping

3.1 Narrower independent claim ("automatic control … involves a turn signal")

Assessment: The '381 maps onto the narrower independent claim essentially element-for-element, including — remarkably — (a) the statistical analysis / learning limitation, (b) adaptive threshold subject matter, and (c) the very "independent of steering-wheel angle" rationale the '696's specification touts as its inventive advantage (¶0016 of the '381 pre-dates and pre-empts it). This is a strong § 102 position on the narrower claim; it is a very strong § 103 position on it even if one limitation (e.g., a construction of "control parameter") is argued to be unmet.

3.2 Broad independent claim ("automatic control of the car" / "function of the car")

Same mapping, minus the turn-signal-specific tying. Note the irony: if the complaints are right that claim 1 recites a "function of the car" rather than "automatic control," the claim becomes broader and easier to invalidate — the '381's lane-position output to a control unit is a "function of the car" on any reasonable construction. Either formulation is met.

3.3 Inferred dependent ladder

---

4. The combinations and the motivation to combine

Ground 1 — The '381 alone (or '381 in view of the knowledge of a PHOSITA): narrow independent claim

Rationale. Anticipation-flavored § 103: the '381 discloses the entire claimed combination, and to the extent any limitation is construed as unmet, the difference is a predictable use of a known technique (lane-position arithmetic → threshold comparison → indicator actuation) to improve a known device of the same kind. KSR, 550 U.S. at 417. Motivation is intrinsic to the reference: the '381 states its own object is to automatically switch the turn indicator on and off, notes that drivers frequently fail to signal, and expressly promotes automatic actuation "regardless of how conscientious the driver is" (¶0011) — i.e., the same problem statement as the '696's Background.

Ground 2 — '381 + US 5,790,403 (+ US 4,970,653): image-processing robustness and inter-frame tracking

Claim features served: "image data from a camera"; the '696's "uses a position of an image of a boundary in a previous image frame to estimate a location of an image of a boundary in a current image frame" (express spec text, and its stated advantage of reduced processing time); curve-fitting to estimate near-field boundary portions from far-field data; template comparison; monitoring a prescribed area of each frame.
Why combine. The '381 delegates its lane recognition to a "lane recognition unit… essentially known" and cross-references DE 195 07 957 C1 (¶0027) — an express, built-in invitation to substitute/upgrade the lane-recognition subsystem. US 5,790,403 addresses precisely the failure mode the '696 identifies ("Sometimes, the processor may not be able to detect the portions… adjacent or relatively closer to the vehicle"): it stores a lane-boundary estimate from a preceding processing cycle as historical information and confines Hough processing to a region derived from it, explicitly to reduce computation. US 4,970,653 (cited as prior art by the '403 itself) does the same for the search window and adds template matching and missing-marker estimation. All three are in the same field (video/laser lane recognition for vehicle control), solve the same problem (real-time lane tracking with dashed, missing, or occluded markers), and produce a predictable improvement. A PHOSITA automating a turn signal from lane data in 2003 would combine them as a matter of ordinary design.

Ground 3 — '381 + US 5,979,581: sensor/estimation alternatives and the "spatial relationship" element

Claim features served: image/sensor data input; determination of the spatial relationship between the car and the lane; output of the control parameter.
Why combine. US 5,979,581 claims a lateral vehicle control system that signals a steering control output from an estimated lateral position differential relative to a lane centerline derived from a retroreflective stripe. It reinforces the well-known, well-documented equivalence of optical-lane-marking sensing approaches (CCD, laser/retroreflective, infrared transit-time) — the '696 itself lists these interchangeably as equivalent sensors, which is an admission that the sensor choice is a design option, not an inventive contribution. Improvement of one element (the sensor/estimator) in a known system, with predictable results, is the paradigm of KSR.

Ground 4 — '381 + US 4,348,652 (and/or '381 ¶¶0039–0041): speed-gated enablement and speed-adjusted criteria

Claim features served: "the processor is configured to receive a speed of the car"; "process the image data and the speed"; and the spec's FIG. 6A/highway-speed apparatus (automatic control only above a prescribed speed; different distance thresholds at different speeds).
Why combine. US 4,348,652 (1982) already teaches inhibiting the automatic road-edge system when vehicle speed is below a preset minimum and when the turn signals are in use — the same "gate" architecture the '696 recites. The '381 supplies speed as a computational input (route length and cyclic advance as a function of speed: ¶¶0039–0041). US 6,580,986 supplies the automatic alteration of timing parameters versus travel speed. Motivation: at low speed (parking, stop-and-go, urban turns) lane-position signals are noise-dominated and nuisance actuation is frequent; disabling or desensitizing automatic lane-change signalling below a speed threshold is a common-sense engineering choice with a predictable benefit. KSR, 550 U.S. at 421.

Ground 5 — '381 + US 6,580,986: speed-dependent sensitivity / adaptive criteria, and the display dependent

Claim features served: "adjust a threshold based on the result from the statistical analysis"; different activation distances at different speeds; the display dependent.
Why combine. US 6,580,986 expressly discloses a lane-marking system in which the governing periods may be altered against variations of travel speed and "the alteration may be automatically carried out against variations of travel speed during travel." The '381 supplies the statistical/neural adaptation layer and the "threshold values" for actuation. Combining an adaptive timing law with a lane-position threshold comparison yields, predictably, exactly the claim-1-dependent "adjust a threshold" behavior. Its display 122 also meets the flat-panel dependent. Any overlap with US 6,580,986 must be date-confirmed (§7, item 3).

Ground 6 — Lever / manual-override dependent

The '381 detects the driver's manual actuation of the indicator (¶¶0048–1053). Every vehicle of the era has a turn-signal lever; the '381's own ¶0005 discusses the DaimlerChrysler "convenient jog control." A "control signal from a lever" is thus either disclosed or a notoriously well-known element whose combination raises no new mechanics. Moreover, the '696's own specification presents the lever-control input (FIGS. 9A/9B) as a user-convenience add-on to an otherwise complete system — the classic "arrange old elements as dictated by their respective functions" case.

Ground 7 (fallback if the AIA applies) — expanded art window

If claim 1 really requires a post-2004 "machine-learning capability," note the tension: the pre-2004 '381 already discloses "neural networks" used for statistical adaptation of the indicator's switching point. So even a "machine learning capability" limitation does not escape the '381 — which is the single most important consequence of verifying the '381 this session. If instead the claims are AIA-era, add the 2004–2016 landscape (e.g., the Continental/Wabco family and the 2016-2018 turn-indicator-automation art that surfaced incidentally: EP 3 330 128 A1; DE 10 2013 219 023 A1; DE 10 2013 225 138 A1; WO 2014/174000 A1; DE 10 2009 047 436 A1; US 9,517,719 B2) — but those are leads only; none has been retrieved in substance, and none may be cited as a ground here.

---

5. Anticipating the patent owner's response

5.1 "The '381 doesn't disclose statistical analysis / learning."
This is the owner's only plausible escape hatch, and it fails: the '381 states the compiled data "is stored and analyzed by statistical methods such as those using neural networks" (¶0051) and that the analyzer "can adapt the point in time… to the driving habits of the driver on the basis of conventional statistical methods such as those using neural networks" (¶0035). Whether the limitation is worded "statistical analysis" (abstract/Definitions) or "machine learning capability" (complaints), ¶0035/¶0051 reach it.

5.2 The "wireless network" dependent is the weak point in the current record.
The '696's dependent claims reciting receipt of information/instructions from a wireless network are supported in the specification only by the FIG. 10 generic-computer disclosure (a remote computer sending instructions over a telephone line/modem → IR, and "the computer system 1200 receives programmed instructions from a wireless network"). The '381's navigation system is satellite/GPS-based (a wireless reception, but a digital map, not "information from a wireless network"), and I did not retrieve US 5,982,278, US 5,957,983, or US 2002/0175813, one of which may supply this. Do not plead this dependent claim as obvious on the present record. Note also US 2005/0131597 (driver-behavior telematics/OnStar broadcast; statistical maneuver recognition) is filed 2004-07-20 — after 2004-03-15 — so it is not prior art if the '696 keeps its provisional date. Flagging this because it looks attractive and is likely a trap.

5.3 Secondary considerations / nexus — and why they are weak here.
The strongest non-obviousness theme the '696 asserts is "independent of a turning angle of the wheels" (spec text: activation is independent of wheel angle, avoiding mistimed actuation on curved lanes). The '381 destroys any nexus: it states verbatim that the turn indicator can be turned off "when changing lanes with a small steering angle, because the basis of this method is the lane change and not the change in the position of the steering wheel" (¶0016). Likewise the '696's "reduce processing time" story is the stated purpose of US 5,790,403's historical-information region-of-interest and US 4,970,653's previous-frame search window. There is no evidence in the prior sections of commercial embodiment, licensing success, or industry praise tied to these claims (the assignee chain is inventors → AutoSignal LLC → Wyoming Technology Licensing LLC, with no verified product).

5.4 "No motivation to combine / teaching away."
Hard to sustain: the '381's lane recognition unit is expressly the known Daimler unit of DE 195 07 957 C1 and can be supplemented or replaced (¶0027); the '381 itself urges using the navigation system in support of lane recognition (¶¶0019–0021, 0057) and adapting to driver style (¶0023). Combining a known turn-signal controller with a known, more robust lane-image processor, a known speed gate, and a known driver-adaptive statistical layer is a combination of references addressing the same field and the same problem, with predictable results — the KSR fact pattern.

5.5 Double-counting caution. If the '381 appears on the '696's or an ancestor's front page (a Google Patents citation table for family member US 9,248,777 B2 lists US20040143381A1 with dates "2002-11-05 / 2004-07-22"), the owner may argue the examiner considered and allowed over it. That is not a defense in litigation — the reference remains available for § 103, and KSR/Graham are applied by the court independent of the examiner — but it does mean the "the PTO never saw this art" narrative may be unavailable and should be confirmed against PatentCenter before being used.

---

6. Bottom line

Overall obviousness posture: strong, and the constraint is claim text, not art.

• On the turn-signal-tied independent claim, the '381 is a genuine § 102 reference, and at minimum supports § 103 alone. It discloses sensor → processor → offset-value/ threshold analysis → data bus → automatic turn-indicator control, plus statistical/neural adaptation, plus driver-style learning, plus automatic deactivation after the lane change.
• On the broad independent claim, the same mapping works and works harder: on the complaints' "function of the car" phrasing it is met outright; and even on an added "machine learning capability" limitation, ¶¶0035/0051 of the '381 supply it.
• The two combinations most likely to carry a § 103 challenge are:
  1. '381 as the primary reference (optionally with US 6,580,986 for automatic speed-based parameter alteration) — for both independent claims and for the threshold-adjustment, speed, display, and lever dependents; and
  2. '381 + US 5,790,403 (+ US 4,970,653) — for the camera-image, inter-frame estimation, curve-fitting, template, and prescribed-area dependents.
     Add US 4,348,652 where a speed-based enable/disable gate is needed.
• Do not plead the "wireless network" dependent on this record; the supporting reference was not retrieved.

---

7. Verification checklist (do these before filing or defending)

1. Pull the '696 granted claims and its front-page "References Cited" (PatentCenter, App. 15/348,766). Resolve (a) "automatic control of the car" vs. "function of the car"; (b) whether claim 1 recites "machine learning capability"; (c) claim numbering; and (d) whether US 2004/0143381 A1 appears on the '696's or an ancestor's face (a citation table for US 9,248,777 B2 lists it — confirm against the actual patents).
2. Determine pre-AIA vs. AIA status by checking written-description support in the 2004/2005 disclosures for every limitation, especially any learning/machine-learning element. This governs the entire art universe.
3. Confirm US 6,580,986's U.S. filing date and its family (EP 1 074 903 B1) — my § 102(e) call is date-dependent and unverified.
4. Retrieve US 5,982,278; US 5,957,983; US 2002/0175813 in full (and US 4,970,653's exact issue date/assignee), and specifically test whether any supplies the wireless-network dependent.
5. Retrieve and date-check US 4,970,653, DE 19921449 C1, JP 2002019520 A, FR 2815311 B1, and US 6,882,287 B2 before using them; and do not import the Unified Patents '916 claim-33 chart into a '696 contention without re-mapping.
6. Do not merge identifiers: DE 10046696 A1 (family member of US 6,629,030 B2) and US 10,469,966 B2 are not this patent. No correction is made to those numbers.