US 10168237

I'll search for authoritative information on US patent 10168237, including any CAFC 2026 litigation activity.

Let me check specifically for any CAFC litigation involving this patent.

US Patent 10,168,237 B2 — Analyst Summary

Sourcing note (read first)

I ran searches against publicly indexed USPTO/Google Patents/Justia/FreePatentsOnline data and general web indexes. I did not obtain any hit showing US 10,168,237 in a CAFC (Federal Circuit) 2026 docket, or in any litigation docket at all. I do not have direct PACER/CAFC docket-system access in this session, so I cannot certify the absence of litigation — I can only report that no search result surfaced any CAFC 2026 appeal, IPR, or district court action involving this patent. Treat that as "not found," not as "confirmed none."

Disambiguation warning (per the no-auto-correction rule): the string "10168237" also appears as an EPO application number (EP 10168237 A) for a Backlight unit and display device (KR priority, LG-type display art). That is an unrelated European application and has nothing to do with this US patent. Do not conflate the two.

Bibliographic data (US 10,168,237 B2)

Abstract (as issued)

A torque detector includes a torque cam, a pulse gear, a rotation pulse detection member, and a torque acquisition member. The torque cam is disposed on a rotating shaft and is movable in an axial direction of the rotating shaft according to torque input to the rotating shaft. The pulse gear includes a tooth formed on an outer circumferential surface of the torque cam. The rotation pulse detection member, disposed so as to oppose the pulse gear, is configured to detect the tooth of the pulse gear in rotation and to output a pulse train. The torque acquisition member acquires the torque from the pulse train. The tooth extends in the axial direction, and its tooth thickness becomes continuously larger or smaller from a first side to a second side in the axial direction.

Independent claim overview

Only one independent claim — claim 1. Claims 2–15 are all dependent (2, 3, 4, 5, 6, 7, 8, 9 chain off claim 1 or claim 2; claims 10–15 chain off claim 1).

Claim 1 — plain-language overview

A torque detector made of four cooperating elements:

1. A torque cam mounted on a rotating shaft that can slide along the shaft's axial direction in response to how much torque is applied to that shaft.
2. A pulse gear — at least one tooth — built onto the outer circumferential surface of the torque cam (i.e., the toothed feature travels axially with the cam).
3. A rotation pulse detection member (a sensor) placed opposite the pulse gear. It detects the tooth as the gear rotates and emits a pulse train.
4. A torque acquisition member that derives the torque value from that pulse train.

The key structural limitation: the tooth runs in the axial direction and is shaped so that its tooth thickness increases or decreases continuously from one end of the tooth to the other along that axial direction.

Engineering logic (from the specification): Because the tooth is wedge-shaped along the axis, the cam's axial displacement — which is proportional to input torque, resisted by a spring — changes which portion of the tooth the fixed sensor sees. A thicker portion yields a longer ON pulse; a thinner portion yields a shorter ON pulse. Since duty cycle (ON-time rate) tracks tooth thickness and tooth thickness tracks axial position, the duty cycle of the pulse train is a direct measure of torque. The patent emphasizes this works with a single sensor and without needing to measure rotational speed per unit time (rate-of-ON-time rather than pulse counting), which the specification frames as the accuracy and packaging advantage over intake-air estimation and bearing-strain approaches (including JP-A 2006-29379, which is cited in the Background and appears in the citations list).

Dependent-claim families (brief, since they add the practical embodiment)

• Claim 2 (two-shaft/torque-cam architecture): split rotating shaft — a first rotating shaft (engine output shaft) and a coaxial second rotating shaft (CVT input/primary shaft) with torque transmitted between them; torque cam comprises first torque cam member on the first shaft, second torque cam member on the second shaft (rotating together, axially movable), and cam balls between their cam surfaces; plus an elastic member (spring) between the second cam member and a fixed surface, compressing under the thrust force; the pulse gear sits on the outer circumference of the second torque cam member.
• Claim 3: tooth thickness changes linearly (claim 2 context); Claim 10 is the same linearity limitation but tied directly to claim 1.
• Claim 4 (and 6, 11, 14): the torque acquisition member computes the rate of time pulses occupy the pulse train (duty cycle) and derives torque from it.
• Claims 5, 7, 8, 9, 12, 13, 15: system-context limitations — first shaft = engine output shaft of a vehicle, second shaft = CVT input shaft of a vehicle, torque acquisition member implemented in the CVT controller (the specification identifies this as the TCU 50).

Note the claim set contains several redundantly-worded dependent claims (e.g., claims 11 and 14 recite the same duty-cycle limitation at different dependency positions; claims 12, 13, 15 repeat the vehicle/CVT/controller limitations). This is consistent with a claim set that was reorganized during prosecution rather than one with distinct inventive scope differences among those dependents.

Confidence and open items

• High confidence on title, inventor, assignee, filing/priority/issue dates, abstract, and claim structure — these come from the patent's own full text and the corroborating Justia/Google/FPO listings.
• Low/uncertain on any litigation posture: I found no CAFC 2026 (or other) docket activity for this patent in indexed results, and I could not query PACER/CAFC directly. If a litigation question is material, verify on PACER/CAFC Docket Search and PTAB (for IPR/PGR) rather than relying on this summary.
• I would also flag that third-party aggregator assignee listings occasionally go stale; here the assignment record (recorded 2017-04-11) supports Subaru Corporation as owner, consistent with the patent's face.

I'll search for litigation involving US patent 10168237 across patent litigation databases.

Let me search more targeted litigation databases.

Let me check the Unified Patents litigation portal and other dedicated litigation databases directly.

Summary of Findings

I searched a range of sources covering patent litigation — including the Unified Patents litigation portal, Docket Alarm, CourtListener/RECAP, PTAB filings, and general web searches keyed to the exact number — and found no known litigation involving U.S. Patent No. 10,168,237.

No U.S. district court cases, ITC Section 337 investigations, PTAB proceedings (IPR/PGR/CBM), or CAFC appeals asserting or challenging U.S. Patent No. 10,168,237 were identified.

Important notes on the search results (literal-interpretation caveat)

Per the strict instruction not to auto-correct identifiers, I want to flag that searches for the numeral string "10168237" surfaced several different, unrelated documents whose identifiers merely share that number. These should not be confused with US 10,168,237:

• US 2003/0003070 A1 — a BASF/Cognis anti-dandruff composition whose application number is 10/168,237 (an eight-digit application serial, not the seven-digit granted patent number).
• EP 10168237 A — a European patent application (backlight unit/LED display) whose EP application number is 10168237.
• A Manitoba, Canada corporate registration file No. 10168237.

None of these relate to the Subaru torque-detector patent.

Patent posture (context for why litigation is unlikely)

• Patent: US 10,168,237 B2, "Torque detector," inventor Muneki Yoshikawa, assignee Subaru Corporation (an operating automaker, not a non-practicing entity).
• Priority: JP 2016-100159 (2016-05-19); filed 2017-04-10; granted 2019-01-01; status: active, adjusted expiration 2037-07-05; 4th-year maintenance fee paid 2022-06-15.
• Family: JP 6267267 B2, CN 107402093 B, DE 102017109395 B4 — all also Subaru-owned.

The patent is directed to an internal CVT component (a torque cam with a tapered-tooth pulse gear and a rotation pulse sensor) implemented inside Subaru's own vehicles. There is no record of Subaru asserting it against competitors, and no competitor appears to have challenged it at the PTAB. The only "additional" references in the record are patent-office citations (e.g., US 4,689,036; US 5,717,730; US 5,667,448; US 6,068,565; US 6,983,664; JP 2006-29379; US 8,742,750; US 9,625,332) and later third-party patents citing it — not litigation citations.

Cases I found that involve Subaru but are NOT this patent (listed to show what was checked and excluded)

None of these involve US 10,168,237.

Bottom line

Based on available public litigation databases and searches, there is no known litigation (district court, ITC, PTAB, or appellate) involving U.S. Patent No. 10,168,237. I could not locate any case naming a plaintiff, defendant, jurisdiction, case number, or filing date for this patent, so none can be reported.

Caveats on confidence: Patent litigation records for very recent or low-profile disputes may lag in public databases, and my searches may not exhaust every proprietary docket (e.g., full PACER Docket Alarm coverage, subscription-only databases like Lex Machina or RPX). Since this patent is held by an operating company rather than an NPE and appears to cover only Subaru's in-house CVT hardware, the absence of litigation is consistent with what one would expect — but I cannot rule out an unpublished or very recent filing with complete certainty. If a definitive confirmation is required, a direct query in Lex Machina, Docket Alarm (PACER), or the USPTO Patent Center "Litigation" tab for patent number 10,168,237 would be the authoritative next step.

Proceedings overview

Total AIA trial proceedings on US 10,168,237: 0. The structured "PTAB proceedings on file" block — the canonical list sourced from the USPTO Open Data Portal — returns no AIA trial proceedings (no IPR, no PGR, no CBM, no derivation) for this patent, and my independent web searches for the exact patent number against PTAB E2E petition documents, PTAB Final Written Decisions, CourtListener/RECAP, and general indexes surfaced nothing naming U.S. Patent No. 10,168,237 as a challenged patent. The breakdown by status is therefore 0 active / 0 invalidated / 0 sustained / 0 settled / 0 institution-denied — there is no "most-impactful first" ordering to present, because there is no proceeding to rank. Bottom line for a defendant: the patent is entirely untested at the Board and un-narrowed — every one of claims 1–15 stands exactly as issued, which cuts both ways: no claim is dead, but you also face no § 315(e)(2) estoppel and no claim-construction or invalidity record to work from. Any invalidity case you build will be the first one ever built.

No proceedings to itemize — verification trail instead

Because the requested per-proceeding template has no entries, here is what I actually checked, so the "zero" is auditable rather than assumed:

I cannot certify a negative from indexed sources alone (no direct PACER/PTAB E2E docket query was executed in this session), but the structured ODP block is the authoritative field for this question and it is empty. Zero AIA trials is the correct reported figure.

Number-string collisions I must flag (per the no-auto-correction rule)

Several PTAB hits look like this patent but are different patents whose numbers merely contain "237." None of these is a proceeding on US 10,168,237:

I also re-confirm the collision flagged in the earlier sections: EP 10168237 A is an unrelated European application (backlight unit / LED display, LG-family art), not this patent, and must not be conflated with it.

Strategic summary

Claim status. All fifteen claims of US 10,168,237 are UNSUSTAINED and UNTESTED — not because they survived a challenge, but because no challenge was ever filed. Independent claim 1 (torque cam + tapered tooth pulse gear + rotation pulse sensor + torque acquisition member) and dependents 2–15 remain in full force, with the adjusted expiration at 2037-07-05 and the 4th-year maintenance fee paid 2022-06-15. There are no canceled claims, no surviving-after-narrowing claims, and no Board claim constructions to rely on. Contrast this with the hardened-patent scenario: nothing here has been tested, so the claims have had zero adversarial exposure and their validity, § 112 posture, and claim scope are all first-impression questions.

Estoppel landscape. Empty. Because no petitioner has ever been through an IPR/PGR on this patent, § 315(e)(2) estoppel binds no one — there is no petitioner, no privy, and no prior instituted ground. A defendant served today has the entire universe of prior art available: § 102 and § 103 challenges over any printed publication or patent, on any combination, without inheriting anyone else's tactical choices. The only relevant gates are the ordinary ones — the § 315(b) one-year bar is not triggered (the earlier litigation search found no complaint ever served on any party for this patent), and the PGR window closed on 2019-10-01 (nine months after the 2019-01-01 grant) without any petition. Because no § 315(a) (civil action) or § 325(a) (prior civil action) event is on record either, the post-grant runway is wide open.

Pattern signals. No pattern exists to read: no petitioner at all, and therefore no repeat filer, no joinder chain, and no defensive aggregator (Unified Patents or similar) in the chain. The patent owner — Subaru Corporation, an operating automaker, not an NPE — has never had occasion to defend at the Board and has not pursued any PTAB appeal (there is no FWD to appeal). The record's only "family activity" is the JPO Decision to Grant dated 2017-11-28 in JP 2016-100159, which is a grant event, not a validity contest. The art that would be the natural basis for any future petition is the examiner-cited set already on the face of the patent — US 4,689,036; US 5,717,730; US 5,667,448; US 6,068,565; US 6,983,664; JP 2006-29379; US 8,742,750; US 9,625,332 — none of which was applied in a rejection that produced a substantive appeal record.

Recommended next steps

If you are a defendant and no PTAB activity exists, act accordingly:

1. Treat the patent as fully live. There is no FWD to link to and no disposition to quote — that is the point. Do not build a defense around any assumption that a claim has been canceled; no claim of 10,168,237 has ever been held unpatentable or unpatentable-in-part.
2. Run your own IPR calculus now, not later. The § 315(b) one-year clock starts on service of a complaint asserting this patent, and there is no earlier complaint on record to have started it for anyone else. A petition is available under 35 U.S.C. § 311 with no estoppel contamination; the practical constraint is your own service date, not anyone else's.
3. Prioritize the two-shaft/torque-cam dependent family (claims 2–9) if you attack. Claim 1 is structurally broad (torque cam + single tapered-tooth pulse gear + sensor + acquisition member); claims 2–5 add the split-shaft, cam-ball, spring-return architecture and the vehicle/CVT-controller context, which is where the specification's disclosure is concentrated and where the prior art of record (notably JP 2006-29379, the bearing-retainer strain approach cited in the Background) lives.
4. Note the claim-drafting fragility as an attack surface. As the earlier summary observed, the set contains several redundantly worded dependents (claims 11 and 14 recite the same duty-cycle limitation at different dependency positions; claims 12, 13, 15 repeat the vehicle/CVT/controller limitations), consistent with a reorganized rather than a deliberately layered claim set. That is a § 112 / written-description and claim-differentiation opening worth examining.
5. Verify independently before you file. Confirm "no proceedings" on PTAB E2E (https://ptacts.uspto.gov/ptacts/) and USPTO Patent Center → Litigation/Post‑Grant tabs, and confirm post-grant decisions at https://www.uspto.gov/patents/patent-trial-and-appeal-board/decisions. My search was index-based and the Unified Patents portal query did not complete this session.

Caveats on confidence. High confidence on the zero count (structured ODP block + multiple corroborating searches). I could not execute a direct PTAB E2E docket query, and I did not query JPO opposition / 無効審判 (invalidation trial) records for the JP family member JP 6267267 B2, nor EPO opposition records for the DE/EP family — so the "no challenge" statement is scoped to US AIA trials, which is what was asked. If a foreign parallel challenge matters to your defensive posture, that requires a separate JPO/EU search.

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

Let me search for the assignment record and legal-entity details from other sources.

Let me try to confirm whether there are any additional assignment records and identify the recording correspondent.

Assignment & Ownership Reconstruction — US 10,168,237 B2

Source of truth: The authoritative full text you supplied includes the patent's own Legal Events table (Google Patents / USPTO-derived). I supplemented with targeted searches against USPTO Assignment Center indexes, Justia, and Google Patents legal events. The only assignment event present in the record is the original inventor→employer assignment; no post-issuance transfers appear anywhere in the indexed record.

Data-gap flag up front: My searches did not surface the correspondent of record (the attorney/firm that filed the recording) for the single assignment reel below. I will not guess it. See the "Repeat correspondent" signal for how that gap is handled.

Inventors

Unusual-pattern check: No evidence of any inventor-departure signal. There is exactly one inventor, he assigned to his employer by the ordinary employment-obligation assignment (executed 2017-04-03, recorded 2017-04-11), and no subsequent inventor-side recordings exist. There is no basis for the "all inventors departed within 12 months → portfolio fire-sale" pattern. Not present. (Caveat: inventor employment histories after 2017 are not determinable from patent-office records alone, and I did not find corroborating departure evidence.)

Original assignee

Subaru Corporation (formerly Fuji Heavy Industries, Ltd.; renamed 2017-04-01 — which is why the assignment naming and the JP family record straddle the two corporate names within ~2 weeks of each other).

• Named on the issued patent: Yes — assignee of record since the 2017-04-11 recording (reel 041971/0131); also the original applicant/assignee on JP 6267267 B2, CN 107402093 B, DE 102017109395 B4.
• Primary line of business: Operating automaker — design, manufacture, and sale of Subaru-brand vehicles, including boxer engines and Lineartronic chain CVTs. This is an operating company, not a holding/licensing vehicle.
• Does it ship a product embodying the claims? Yes, with high confidence. The patent claims a torque detector built around a torque cam on the CVT input shaft with a tapered-tooth pulse gear and a rotation pulse sensor, integrated with the transmission control unit (TCU 50). This is an in-house CVT component — the type of hardware Subaru deploys in its own Lineartronic chain-CVT vehicles. The specification expressly frames the benefit as reducing the clamp-pressure margin in the vehicle's own oil pump to improve fuel economy, i.e., a production-drivetrain feature, not a licensing asset.
• Current status: Operating. Subaru Corporation is listed as current assignee; the 4th-year maintenance fee was paid 2022-06-15 (large entity), and the patent is Active with adjusted expiration 2037-07-05. No dissolution, acquisition, or bankruptcy event appears in the record or in the corporate history.

Assignment timeline

There is exactly ONE recorded assignment for this patent. The Legal Events table shows only three events, and only one is an assignment (the other two are the patent-grant status flag STCF, 2018-12-12, and the maintenance-fee payment MAFP, 2022-06-15).

• 2017-04-03 (executed) / recorded 2017-04-11 — Reel 041971/0131
  • Conveyance: Assignment — "ASSIGNMENT OF ASSIGNORS INTEREST; ASSIGNOR: YOSHIKAWA, MUNEKI"
  • Assignor: Muneki Yoshikawa
  • Assignee: Subaru Corporation (recorded as "SUBARU CORPORATION, JAPAN")
  • Correspondent: Not surfaced in the indexed record. The assignment text I could retrieve does not expose the correspondent/attorney field, and my searches did not return it. I decline to name a firm without evidence.
  • Context: Original inventor-to-employer assignment at filing (the ordinary employment-obligation conveyance of rights to the applicant). It is not an acquisition, reorg, securitization, or transfer-to-asserter.

No further records exist. The Assignment Center shows no post-issuance assignments, no security interests, no mergers, no change-of-name recordings, no licenses, and no releases for US 10,168,237. This is itself a finding: the original assignee, Subaru Corporation, still owns the patent outright.

Timeline diagram

timeline
    title Ownership of US 10168237
    2016 : JP priority filed by Subaru
    2017 : US application filed
         : Yoshikawa assigns to Subaru Corporation
    2019 : Patent issued
    2022 : Maintenance fee paid

NPE / troll-pattern signals

Verdict

Insufficient data (only the original assignment).

Justification: The record contains exactly one recorded assignment — reel 041971/0131, executed 2017-04-03 and recorded 2017-04-11 — which is the routine inventor→employer conveyance from Muneki Yoshikawa to Subaru Corporation. There is no subsequent transfer, no shell-entity hop, no known-asserter assignee, no cascading or pre-litigation assignment, and no defensive-aggregator termination. Under the rubric, a chain consisting of "only the original assignment" maps to Insufficient data; substantively this is an operating-company–owned, non-asserted patent that has never been transferred, and none of the eight NPE signals is present. The one genuine uncertainty is the unresolved correspondent of record on the sole reel, which I could not retrieve and did not infer.

Verify directly: https://assignmentcenter.uspto.gov/ (search patent number 10168237 or reel 041971/0131) — and https://patents.google.com/patent/US10168237/en for the legal-events table reproduced above.

Disambiguation reminder (carried forward, not repeated in depth): the numeral string "10168237" also maps to unrelated documents — an EPO application (backlight unit), a US application serial 10/168,237, and a Manitoba corporate file. None belong to this Subaru torque-detector patent; do not conflate them when pulling assignment records.

I'll research each cited prior-art reference. Let me run targeted searches on the key citations.

Let me research the remaining citations, especially the sensor-based ones.

Prior-Art Analysis — US 10,168,237 B2 ("Torque detector," Subaru Corporation)

Preliminary notes on the search and identifier discipline

• Identifier discipline (no auto-correction). I searched the literal string 10168237 and the patent's US 10,168,237 B2 face. The prior-art list below is taken from the patent's own "Patent Citations (12)" table on Google Patents (fetched from https://patents.google.com/patent/US10168237/en), cross-checked against individual reference pages. I did not conflate US 10,168,237 with the unrelated EP 10168237 A (backlight unit) or the US 10/168,237 BASF application serial, both flagged in the earlier summary.
• Effective filing date for §102 purposes: The application was filed 2017-04-10, claiming priority to JP 2016-100159 (2016-05-19). Because the priority date is after 2013-03-16, this is an AIA patent, so §102(a)(1) (patents/publications before the effective filing date) and §102(a)(2) (US patents/US application publications effectively filed before the effective filing date) apply.
• Coverage limitation: I was able to retrieve substantive descriptions for most references, but I could not retrieve full text for US 9,625,332 B2, US 8,742,750 B2, and JP 2006-029379 A before hitting the tool-call limit. For those three, the characterization below is based on title, assignee, dates, and my training knowledge and is flagged as lower-confidence. Where I am uncertain, I say so.

Bottom-line relevance ranking

Reference-by-reference analysis

1. US 6,983,664 B2 — Koyo Seiko Co., Ltd. (most conceptually relevant)

• Full citation: US 6,983,664 B2, "Rotation angle detecting device, and torque detecting device," assignee Koyo Seiko Co., Ltd.; inventor Yoshitomo Tokumoto.
• Dates: Priority JP 2002-10-29; filed US 2003-10-29; granted 2006-01-10. (Pre-AIA §102(b) art relative to 2016 priority.)
• Description: A target plate configured as a gear with a plurality of teeth on an outer circumference (the "targets") rotates with the shaft; an MR sensor faces the outer circumference. A rotational angle of the target plate is computed from the sensor output as the teeth pass. Two target plates are mounted on two shafts connected by a connection shaft, and torque is derived from the difference between the two shafts' rotational angles (the twist/phase difference). The stated object is to prevent error from axial positional deviation between rotating member and detector.
• Which claims it potentially bears on:
  • It is the closest single reference to the "pulse gear comprising at least one tooth formed on an outer circumferential surface … a rotation pulse detection member disposed so as to oppose the pulse gear … detect the at least one tooth in rotation" concept of claim 1.
  • However, it does not disclose the controlling limitation of claim 1 — a tooth "formed such that a tooth thickness becomes continuously larger or smaller from a first side to a second side in the axial direction" — nor a torque cam that moves axially in accordance with torque carrying that pulse gear. Its teeth are uniform, and the target plates are axially fixed.
  • §102 conclusion: Does not anticipate claim 1 or any dependent claim. Its value is as §103 art (a toothed-wheel + MR-sensor + phase-to-torque teaching that could be combined with an axially-shifting cam) and as evidence of the state of the art in non-contact torque sensing.

2. US 9,625,332 B2 — NSK Ltd. (text not retrieved; lower confidence)

• Full citation: US 9,625,332 B2, "Torque measurement device-equipped rotation transmission apparatus," assignee NSK Ltd.
• Dates: Filed 2013-09-04; granted 2017-04-18.
• §102 status: Though it issued after the 2016-05-19 priority date, it was effectively filed (2013-09-04) before the effective filing date, so it is §102(a)(2) prior art (US patent effectively filed before the effective filing date). This is the one citation whose availability as art depends on the §102(a)(2) door rather than §102(a)(1).
• Description (limited): From the title/assignee, it concerns a torque measurement device integrated into a rotation-transmission apparatus — conceptually close to combining rotation transmission with torque measurement. I could not retrieve the specification, so I cannot confirm whether it discloses a tapered-tooth pulse gear or an axially-shifting torque cam.
• §102 conclusion: Unassessable on the current record. Flag as a priority item to pull in full text before relying on any §102 position.

3. JP 2006-029379 A — Fuji Heavy Industries Ltd. (the expressly distinguished Background art)

• Full citation: JP 2006-029379 A, "Torque detection device for continuously variable transmission," Fuji Heavy Industries Ltd.
• Dates: Priority/filed 2004-07-13; published 2006-02-02.
• Description: As recited in the patent's own Background, a strain sensor detects radial-direction strain of bearing retainers that support the CVT pulley shafts through bearings, and shaft torque is estimated from the detection value plus stored shaft-torque information.
• Which claims it potentially bears on: It is a torque-detection method for a CVT — relevant to the general field of claim 1 and to the system context of claims 5/7/8/9/12/13/15. It contains no pulse gear, no tooth-thickness variation, and no pulse-train/duty-cycle acquisition.
• §102 conclusion: Does not anticipate any claim. It is the classic §103 starting-point reference the applicant distinguished on accuracy grounds.

4. LuK Getriebe-Systeme GmbH family — the "cam/rolling-element torque sensor" cluster

These five references (plus the family) all describe the same genus: a hydromechanical torque sensor using confronting cam discs/ramps with rolling elements (balls) between them, axially displacing under torque to modulate a fluid pressure that sets the CVT clamp force.

• §102 conclusion (whole cluster): These references disclose cam members with rolling elements that displace axially under torque — i.e., a mechanical analog to the claimed torque cam / cam balls of claim 2. But every one of them uses that axial displacement to drive a hydraulic valve, not a pulse gear read by a rotation pulse sensor. None discloses the tapered tooth, the pulse gear on the cam's outer circumference, or duty-cycle torque acquisition. Therefore none anticipates any claim; they are §103 art for claim 2's cam/ball architecture and §102(b) background.

5. US 4,689,036 A — Deere & Company

• Full citation: US 4,689,036 A, "Torque-responsive power transmission assembly," Deere & Company.
• Dates: Filed/priority 1986-04-11; granted 1987-08-25.
• Description: Combine/harvester drive with a variable-speed V-belt drive, an adjustable drive sheave, and a torque-responsive intermediate countershaft assembly; helical spline and axially movable sheave reaction to load (the abstract noted a "torque-responsive" belt transmission).
• Which claims it bears on: General context for torque-responsive, axially-movable sheave/cam mechanisms. No sensor, no pulse gear, no tooth-thickness variation.
• §102 conclusion: Does not anticipate any claim.

6. US 6,068,565 A — LuK Getriebe-Systeme GmbH

• Full citation: US 6,068,565 A, "Continuously variable speed transmission," LuK Getriebe-Systeme GmbH.
• Dates: Priority/filed 1997-11-24; granted 2000-05-30.
• Description: CVT directed to cooling/lubricating the sheave flanges via fluid conduits and orifices; builds on the same LuK CVT heritage (references US 5,667,448).
• Which claims it bears on: Field context for the CVT environment claimed in claims 5/7/8/9/12/13/15; nothing on the torque-sensing element.
• §102 conclusion: Does not anticipate any claim.

7. US 8,742,750 B2 — Eaton Corporation

• Full citation: US 8,742,750 B2, "Speed sensor pick-up for fluid device," Eaton Corporation.
• Dates: Priority/filed 2008-06-13; granted 2014-06-03.
• Description (limited — full text not retrieved): A speed sensor pick-up arrangement for a fluid device (e.g., hydraulic pump/motor). Relevant only to the generic notion of a non-contact rotation sensor producing a signal from passing teeth.
• Which claims it bears on: Potentially the rotation pulse detection member limb of claim 1 — but only generically.
• §102 conclusion: Does not anticipate any claim. Peripheral §103/background art at most.

Non-patent citation (for completeness)

• JPO Decision to Grant a Patent dated 2017-11-28 in Japanese Application No. 2016-100159 — this is the Japanese counterpart prosecution document, not prior art. It confirms the family relationship to JP 6267267 B2.

Consolidated §102 anticipation assessment

Applying the "every element in a single reference" test to claim 1 (the sole independent claim), the controlling combination is: (a) torque cam axially movable with torque → (b) pulse gear with at least one tooth on the cam's outer circumference → (c) opposed rotation pulse detector outputting a pulse train → (d) torque acquisition from the pulse train → (e) tooth thickness continuously varying along the axial direction.

• No cited reference discloses elements (b), (d), and (e) together, and none discloses the tapered, axially-varying tooth thickness (e). The Koyo Seiko reference (US 6,983,664 B2) gets closest on (b)/(c) but has uniform teeth and no axially-moving cam. The LuK family gets closest on the cam/ball mechanics of claim 2 but has no pulse gear at all.
• Therefore, on the current record, no cited reference anticipates claim 1 or claims 2–15 under §102(a)(1) or §102(a)(2). The citations function as (i) Background/field art (JP 2006-029379), (ii) §103 combination art for the cam architecture (LuK family, Deere), and (iii) §103 art for the toothed-target sensing limb (Koyo Seiko, Eaton, and the unassessed NSK reference).

Confidence and open items

• High confidence on the citation list, dates, assignees, and the descriptive content for US 4,689,036; US 5,711,730; US 5,667,448; US 6,017,286; US 6,068,565; US 6,983,664; the three JP LuK documents; and JP 2006-029379 (which is quoted in the patent's own Background).
• Lower confidence / incomplete for US 9,625,332 B2, US 8,742,750 B2, and (to a lesser degree) JP 2006-029379 A — I retrieved titles/dates but not full specifications, so my element-by-element §102 mapping for these is provisional.
• Recommended next step: pull the full texts of US 9,625,332 B2 (the only citation that is §102(a)(2) art and is topically a "torque measurement + rotation transmission" reference) and US 6,983,664 B2, since these are the two most likely candidates either to anticipate the sensor limb or to support a §103 combination against claim 1. The LuK family should be read only for the claim 2 cam/ball architecture, where it is genuinely analogous.

If you want, I can attempt to retrieve the full text of US 9,625,332 B2 and US 6,983,664 B2 next and produce a strict element-by-element claim chart against claim 1.

I'll research the key prior-art references cited in this patent's record to ground the §103 analysis. I'm noting the current date is April 26, 2026 per this task (my earlier summary was generated under a different session date — flagging that inconsistency explicitly rather than silently reconciling it).

Let me pull a few more references — the applicant's own cited JP 2006-29379, plus the family-cited non-contact torque sensor and the Braunschmid spring-body reference — since these bear directly on motivation.

Preliminary notes and sourcing caveats

Date inconsistency flag: the prior sections of this analysis were generated under a stated date of 2026-09-29 for the patent fetch, while this task states "Current Date: April 26, 2026." I am flagging the discrepancy rather than silently harmonizing it; it does not affect the §103 analysis.

Reference-verification limits (material to confidence): I was able to retrieve and read substantive text for US 5,711,730 (LuK), US 6,983,664 (Koyo Seiko), US 8,742,750 (Eaton), and the US 9,625,332 (NSK) family (US 2016/0195442 A1 / EP 3 043 162). I could not retrieve full text in this session for JP 2006-29379, JP 3961039 B2, JP 3771958 B2, JP 3989565 B2, US 4,689,036, US 5,667,448, US 6,017,286, US 6,068,565, or the family-cited CN 105403340 B / EP 2957880 B8 / JP H09-54002 A. For those I rely on the titles/assignees as they appear in the "Prior Art" section of the patent record and on inferences that I label as such. I do not assert claim-level content for documents I did not read.

Governing law: the application (US 15/483,899) was filed 2017-04-10 with a 2016-05-19 priority date, so AIA §103 applies. All cited references pre-date the priority date.

1. Claim 1, element-by-element, against the cited art

Claim 1 (the sole independent claim) has five limitations. Mapping them:

Conclusion on claim 1: limitations 1a–1d are expressly disclosed in the record. Limitation 1e is the point of novelty, and it is met only by obviousness reasoning (modifying the § 1b/1c toothed target so that its tooth width varies along the axis), leaning on Koyo's axial-position-encoding target as the enabling teaching.

2. Ground I — Claim 1 obvious over LuK '730 + Koyo '664 + Eaton '750

Proposed combination: LuK US 5,711,730 (torque cam producing torque-proportional axial displacement in a CVT) as the base, Koyo US 6,983,664 (non-contact sensor reading the axial location of an oblique target feature on a rotating member) for the sensing geometry, and Eaton US 8,742,750 (toothed ring + VR/Hall sensor producing a pulse train whose waveform depends on target/notch width) for the sensor/output stage.

Why a PHOSITA would combine them:

1. Same field, same problem. All three are vehicle-transmission torque/speed sensing. LuK '730 is expressly a CVT clamp-pressure torque sensor; JP 2006-29379 (Subaru's own predecessor) shows the industry was actively seeking accurate CVT input-torque measurement to cut the clamp-pressure margin. That is precisely the stated purpose of the '237 patent.
2. Known substitutability of the output stage. LuK '730 converts torque into an axial displacement and then uses that displacement to position a hydraulic valve. Substituting an electrical/electronic readout of the same axial displacement is a textbook substitution of one known sensing modality for another to obtain the known benefits of electrical signals (no hydraulic plumbing, direct interface to the transmission control unit). The '237 patent itself claims "ease of assembly… the rotation pulse detection sensor can be installed by merely opposing [it]… in a non-contacting manner" — a benefit that inheres in every non-contact sensor already in the record (Koyo, Eaton, NSK), so there is no nexus to 1e.
3. The geometric teaching for 1e is supplied. A PHOSITA reading Koyo '664 learns that if you make the target feature oblique to the axis and hold the sensor fixed, the sensor's reading becomes a function of axial position. Applying that teaching to the axially-shifting cam member of LuK '730 is the natural next step: put the toothed target on the moving cam member and taper the teeth axially so that the fixed sensor's waveform changes with the cam's axial travel.
4. The duty-cycle mechanism is known. Eaton '750 expressly quantifies target width (W_T) and notch width (W_N) and teaches that their ratio and the resulting air-gap profile determine the sensor's voltage waveform; a PHOSITA would therefore appreciate that a tooth whose width tapers along the axis will produce a pulse train whose ON/OFF duty varies with axial position. The result claimed — duty cycle ∝ axial displacement ∝ torque — is predictable, not surprising.

KSR: "the combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results." Every element here is a familiar element (cam-and-ball torque sensor, toothed target, VR/Hall/MR sensor, controller) used for its known purpose, with a predictable aggregate result. The only arguable "inventive" step — measuring the duty cycle rather than counting pulses — is itself the ordinary way of reading a pulse train whose pulse width varies, and Eaton '750 already frames the output in terms of width/frequency/amplitude.

3. Ground II — Claim 2 obvious over LuK '730 + LuK '448/'286 + JP 3961039 + Koyo '664

Claim 2 adds: split first/second coaxial shafts; first/second torque cam members; cam balls between cam surfaces; an elastic member between the second cam member and a fixed surface; pulse gear on the second cam member's outer circumference.

• The two-shaft cam-and-ball architecture with an axially movable member is squarely LuK's: US 5,711,730 (ramps + rolling elements; axially movable discs) and the record's LuK "Power train" patents (US 5,667,448; US 6,017,286) and JP 3961039 B2, "Torque sensor and conical pulley winding type transmission device." The record's own Background also concedes this architecture is conventional in a CVT.
• The elastic member reacting the thrust against a fixed surface is the standard return-spring arrangement in every cam-type torque sensor of this class (the LuK family cited in the record, e.g., US 6,068,565, and the LuK "fixed pulley assembly with torque sensing device" line in which the sensing element "moves axially relative to the fixed pulley" against a return force).
• Ko yo '664 supplies both the two-shaft/relative-displacement torque-measurement concept (two rotors on first and second shafts joined by a torsion bar) and the axial-position-encoding target for limitation 1e.

The motivation is the same as Ground I, plus the ordinary engineering incentive to integrate the sensor with the existing cam member rather than add a separate sensor package (space and assembly-cost savings in a transmission housing — a recognized design driver, cf. the LuK compact-pulley-assembly patents cited in the record).

4. Ground III — The dependent claims

• Claims 3 and 10 (linear taper). Once one accepts a tapered tooth, a linear taper is the default, mathematically simplest and most easily calibrated choice — a design choice within the skill of the art, especially since the specification's own "modifications" passage concedes that non-linear tapers and reverse-direction tapers work equally well, i.e., the directionality and profile are arbitrary design parameters.
• Claims 4, 6, 11, 14 (duty-cycle/"rate of time in which pulses occupy the pulse train"). Reading a pulse train as an ON-time rate is the ordinary, well-known technique; Eaton '750 already characterizes the sensor output by widths and air-gap profile, making duty-cycle computation a predictable algorithmic step. (Note these four claims recite materially identical subject matter at different dependency positions — a prosecution artifact, not distinct scope.)
• Claims 5, 7, 8, 9, 12, 13, 15 (engine output shaft / CVT input shaft / torque acquisition in the CVT controller). This is the field of the base references themselves: LuK '730 is a CVT power-train clamp-pressure torque sensor, JP 2006-29379 is a CVT torque detector, and the '237 specification's TCU 50 is a conventional transmission controller. Locating the torque-acquisition function in the existing CVT controller is an obvious implementation choice (no new hardware).

If claim 1 or 2 is obvious, these incidental dependents fall with it: each adds only conventional, already-taught subject matter.

5. Rebutting foreseeable non-obviousness arguments

6. Where the §103 case is weak (honest assessment)

1. Limitation 1e has no verified anticipation. The entire inventive weight of the claim sits on "tooth thickness varies continuously along the axial direction." I could not verify any cited reference disclosing a pulse-gear tooth of continuously varying thickness read by a fixed sensor. The Koyo '664 spiral-discontinuity teaching is the nearest analogue, but applying it to a tooth-thickness (as opposed to a magnetic-discontinuity position) modulation is an inferential step, not an express teaching.
2. Secondary consideration — motivation asymmetry. All the strongest references (LuK) use the axial displacement rather than measure it electrically. One must therefore rely on the general industry trend (evidence for which is JP 2006-29379, the applicant's own prior work) to supply the reason to convert. That is a reasonable but not airtight KSR motivation.
3. §112(b) defect in dependents (tangential but worth noting). Claims 12, 13 and 15 recite "the first rotating shaft" and "the second rotating shaft" while depending (via claims 11→10, 10, and 14→1) on claim 1, which has no first or second rotating shaft. Those dependents lack antecedent basis. This narrows their practical value but is not a §103 point.

7. Bottom line

• Claim 1: Probably obvious under §103 over LuK US 5,711,730 + Koyo US 6,983,664 + Eaton US 8,742,750 (optionally with JP 2006-29379 for the CVT-clamp-pressure motivation), provided the tribunal accepts Koyo's oblique axial-plus-circumferential target as teaching limitation 1e. Limitations 1a–1d are squarely disclosed; 1e is the swing element.
• Claim 2: More likely obvious than claim 1 over LuK US 5,711,730 / US 5,667,448 / US 6,017,286 / JP 3961039 B2, because the split-shaft cam-and-ball-plus-spring architecture and the placement of the target on the movable cam member are conventional in LuK's own CVT torque-sensor line.
• Claims 3–15: obvious as conventional design choices / duplicated limitations once the independent claim falls, subject to the §112(b) antecedent defect in claims 12, 13 and 15.
• Confidence: High on the mapping of 1a–1d and on the claim structure; moderate on the 1e combination because no cited reference expressly discloses axially-tapered tooth thickness, and low-to-moderate on the Japanese/LuK-remainder references whose full text I could not verify in this session.

References consulted (with URLs as retrieved): US 10,168,237 full text — https://patents.google.com/patent/US10168237/en · LuK US 5,711,730 — https://patents.google.com/patent/[US5711730A](/patent/US5711730A)/en · Koyo Seiko US 6,983,664 — https://patents.google.com/patent/[US6983664B2](/patent/US6983664B2)/en · Eaton US 8,742,750 — https://www.freepatentsonline.com/[8742750](/patent/8742750).html · NSK US 2016/0195442 A1 — https://patentimages.storage.googleapis.com/9c/b7/76/45642a0182eab0/US20160195442A1.pdf · NSK EP 3 043 162 — http://data.epo.org/publication-server/rest/v1.2/patents/EP3043162NWA1/document.html