US 11131552

I'll verify current records for this exact patent number and check for any 2026 appellate activity.

Verification note

I ran web searches against public sources (Google Patents records, USPTO-hosted documents surfaced via search, and third-party litigation databases). I could not find any Court of Appeals for the Federal Circuit docket activity — in 2026 or any other year — involving patent 11,131,552. No district court, ITC, or PTAB proceeding asserting this patent surfaced either. I do not have direct query access to PACER or the CAFC's own docket system through these tools, so this is a negative search result, not a certified clearance. Treat "no litigation found" as uncertain rather than as a definitive all-clear.

Search results that did return for this exact number consistently confirm the bibliographic data below and do not contradict the full text. I did not see results for similarly-numbered patents, per your instruction.

Bibliographic data — US 11,131,552 B2

Assignment-record detail worth preserving literally: the initial recorded assignment named "TOYOTA JIDOSHA KABUSHIK KAISHA." A 2021-08-12 corrective assignment corrected the first assignee's name to "TOYOTA JIDOSHA KABUSHIKI KAISHA." So both spellings appear in the chain of title; the corrected form is the operative one.

Abstract (as granted)

"A map generation system includes a server configured to: transfer reference curved road constituent points onto a measurement target curved road as virtual constituent points arranged along the measurement target curved road adjacent to the reference curved road; acquire virtual trajectory curvature information based on positions of the virtual constituent points; acquire detailed curve information of the measurement target curved road by acquiring, based on a position of a start point of the measurement target curved road preset in the road map information, travel information of a vehicle traveling on the measurement target curved road, and the virtual trajectory curvature information, measurement target curved road constituent points arranged at preset intervals from the start point along the measurement target curved road, and associating the virtual trajectory curvature information with each of the measurement target curved road constituent points; and generate a map by acquiring the second detailed curve information."

The problem and the technical approach

Road maps of the "LeanMap" type store curved roads as a start point plus constituent points spaced at preset offsets, each carrying shape information (position coordinates and an azimuth angle θ). "Deepening" means refining that shape information — e.g., updating azimuth angles using curvature radius obtained by dividing vehicle speed by yaw rate over repeated passes.

The disclosure's premise: one curved road (the reference curved road) may already be deepened, while a nearby parallel or opposite road (the measurement target curved road) is not. Rather than accumulating a fresh, independent set of measurement drives on the target road, the system transfers the reference road's already-validated shape information across the lane mark and reuses it.

The mechanics:

1. Transfer (constituent point transfer unit 14) — each reference point P_i (x_i, y_i, θ_i) is projected onto the target road as a virtual point Q_i using X_i = x_i + W·cos(θ_i − π/2), Y_i = y_i + W·sin(θ_i − π/2), where W = (W₁ + W₂)/2, the average of the two lane widths.
2. Virtual trajectory (unit 15) — for adjacent Q_i, Q_{i+1}, the curvature center O_i is the intersection of lines P_iQ_i and P_{i+1}Q_{i+1}; the curvature radius r_i is the distance from O_i to Q_i/Q_{i+1}. An arc of radius r_i about O_i is populated with virtual waypoints R_{i,j} (e.g., every 5 cm), chained into a virtual trajectory L_x.
3. Map generation (unit 16) — r_i is associated with whichever target-road constituent point falls between Q_i and Q_{i+1}, using the vehicle's travel distance from the preset target-road start point. The patent expressly notes Q_i need not coincide with a constituent point — the temporary virtual trajectory bridges that mismatch.
4. When accuracy crosses a threshold, the target road's detailed curve information is promoted into the road map database as FIX data for autonomous-driving and ADAS control (ACC, LKA).

The stated benefit: fewer measurement passes on the target road are needed to reach a given accuracy than if the reference road's detailed curve information went unused.

Plain-language overview of the independent claim

Claim 1 — the only independent claim. A map generation system comprising a server that:

1. Transfers multiple pre-set reference curved road constituent points onto an adjacent measurement target curved road, creating virtual constituent points arranged along it. The reference points are preset in road map information, are associated with the reference road's shape information, and the road map information contains "first detailed curve information" that includes that shape information.
2. Acquires virtual trajectory curvature information — i.e., curvature of a virtual trajectory running along the measurement target curved road.
3. Sets the virtual trajectory based on the positions of the virtual constituent points and the shape information of the reference constituent points that correspond to them.
4. Acquires "second detailed curve information" of the measurement target road by (a) obtaining measurement-target constituent points spaced at preset intervals from a preset start point, and (b) associating the virtual trajectory curvature information with each of them. This is grounded in three inputs: the preset start point position in the road map information, travel information of a vehicle traveling the target road, and the virtual trajectory curvature information.
5. Generates a map by acquiring that second detailed curve information, "for autonomously controlling an autonomous vehicle."

Analyst's observations on claim 1 as literally drafted: the final limitation is phrased as a purpose/use statement ("for autonomously controlling an autonomous vehicle") rather than as a concrete server action, so it may carry little separate patentable weight in an infringement analysis, though it does color the field of use. Claim 1 also mixes method-like steps ("transfer," "acquire," "set," "generate") with an apparatus preamble ("a map generation system comprising a server configured to"), which is common but worth flagging if claim scope is ever litigated.

Dependent claims at a glance

Claim 2 — Adds the geometric specifics: transferring across the lane mark dividing the two roads; obtaining the virtual trajectory's curvature center by intersecting (i) the line through a first virtual point and its corresponding reference point, and (ii) the line through an adjacent second virtual point and its corresponding reference point; and deriving the curvature radius as the distance from that center to adjacent virtual points. This is essentially the O_i / r_i construction described in the specification.

Claim 3 — Adds camera-based refinement: estimating the lane-width-direction center line of the measurement target road from a captured image of a camera on a vehicle traveling that road, and resetting the measurement target constituent points onto that center line. This addresses lateral placement accuracy, complementing the along-road accuracy of claims 1–2.

Claim 4 — A purely geometric relational limitation: an arc representing a portion of the virtual trajectory between two adjacent virtual constituent points shares a center point with an arc extending through the two corresponding reference curved road constituent points. In other words, the transferred target-road arc and the source reference-road arc are held to be concentric (at O_i).

Note the shift in scope: claims 2–4 are all dependent on claim 1, so each incorporates claim 1's limitations plus the additional features. Claim 4 is the most geometrically restrictive; claim 3 is the most hardware/imaging-specific; claim 2 sits in between and is the most likely to be asserted against a comparable cross-lane-transfer implementation.

Uncertainty and limitations of this report

• Docket search: I found no litigation, ITC, or CAFC record for 11,131,552, in 2026 or otherwise. I cannot verify this authoritatively without direct PACER/CAFC access, so do not treat it as a complete litigation-clearance result.
• Claim annotations (the observations on claim 1's means-plus-function-style phrasing and purposes) are my analytical reading, not adjudicated claim constructions.
• No auto-correction applied: the patent number, application number, publication numbers, and the two spellings of the Toyota assignee name appear above exactly as the records render them.

I'll search for litigation involving US patent 11,131,552.

The initial searches returned irrelevant results. Let me search more targeted sources.

Let me search additional litigation-specific databases.

Litigation Search Results: US Patent No. 11,131,552

Bottom line: I found no evidence of any litigation involving US Patent No. 11,131,552.

Based on my searches of publicly indexed sources (general web, litigation-oriented content such as PTAB/ITC filings, CourtListener-hosted documents, and third-party patent-litigation commentary), I could not identify any district court action, ITC investigation, PTAB proceeding, or Federal Circuit appeal in which US 11,131,552 was asserted, challenged, or otherwise at issue. I am therefore unable to populate any of the requested fields (plaintiff, defendant, jurisdiction, case number, filing date, outcome/status) because there is no case to report.

Important caveats about this negative finding

1. This is not a definitive "no litigation exists" conclusion. My searches were largely keyword-based and did not return any hits that actually matched this patent. I was unable to successfully retrieve results directly from the specific databases you named (e.g., the Unified Patents litigation portal, PACER/RECAP via CourtListener) — my queries to those sources either returned unrelated material or were cut off before returning results. A proper clearance-style check would require querying those databases directly by patent number.

2. Search noise to be aware of. A search on the bare string "11131552" returns many false positives. For example, it matches UK Companies House number 11131552 for an unrelated entity ("RUZFISHERMONGER LIMITED," a London discount store in creditors' voluntary liquidation). It also surfaced a genealogy record and various unrelated court documents. None of these relate to the patent. If you or another reviewer run this check, be careful not to conflate such hits with the patent.

3. Nothing I found contradicts the patent record. The Google Patents page for US 11,131,552 lists a "Families Citing this family" section with two later Chinese publications (CN112050821B and CN113932794B) and a "Citations (22)" prior-art list, but it contains no litigation or "Litigation" section — consistent with there being no known litigation.

Context relevant to assessing litigation risk

• Patent: US 11,131,552 B2, "Map generation system."
• Application: US 16/705,682; filed 2019-12-06; granted 2021-09-28.
• Priority: 2018-12-10 (JP 2018-231019; corresponding JP 7060496B2).
• Current/Original assignees: Toyota Jidosha Kabushiki Kaisha and The University of Tokyo.
• Claim 1 is directed to a map generation server that transfers reference curved-road constituent points onto a measurement target curved road as virtual constituent points, acquires virtual trajectory curvature information, derives "second detailed curve information," and "generate[s] a map by acquiring the second detailed curve information for autonomously controlling an autonomous vehicle."
• The assignee profile (a major automaker plus a university) is the type of entity that could assert such a patent, but that is a strategic observation only — I found no record of it doing so.

Recommendation

To reach a legally reliable answer, directly query the following by the exact number (11,131,552 / US11131552 / 16/705,682) rather than by keyword:

• Unified Patents Litigation Portal (https://portal.unifiedpatents.com/litigation)
• PACER / CourtListener RECAP (patent docket search)
• USPTO Patent Trial and Appeal Board (IPR/PGR/CBM by patent number)
• USITC EDIS (Section 337 investigations)
• CAFC docket search

If you'd like, and if I'm given a fresh allowance of search steps, I can attempt these specific database lookups by patent number and report back. As of my searches to date, however, the accurate answer is that no litigation involving US 11,131,552 is known.

PTAB Proceedings on US11131552 — Verification Report

Verification date: 2026-09-29
Sources checked: USPTO ODP structured "PTAB proceedings on file" block (supplied in prompt); targeted web searches for IPR/PGR/CBM petitions and Federal Circuit activity naming US 11,131,552 and its family (JP 7060496B2 / JP 2018-231019); Google Patents prosecution and citation records for US11131552B2.

Confidence statement: The ODP structured data is the canonical source here and it returns zero AIA trial proceedings. My supplemental web searches returned no petition, institution decision, FWD, or CAFC appeal naming this patent. I did not locate any proceeding number, so I am not reporting any — inventing an IPR number would be worse than reporting the null. The one caveat I will flag honestly: negative search results are weaker evidence than a positive hit, and the ODP ingest can lag newly-filed petitions by weeks. If a petition was filed in the last ~30 days, it may not yet be indexed.

Proceedings overview

Total AIA trial proceedings on US 11,131,552: 0. There is no PTAB activity to break down — no active trials, no claims invalidated, no claims sustained, no settlements, no institution denials.

Bottom-line defensive posture for a defendant: You get no free win and no estoppel shield from the PTAB record. This is not a patent that has been hardened by surviving inter partes review, and it is not a patent with canceled claims you can point to. It is an untested patent: all four claims (1–4) are presumptively valid, never subjected to an AIA challenge, and — critically — never construed by the Board. For a defendant, that cuts both ways: nothing is dead, but nothing has been affirmed either, and the entire § 102/§ 103 prior-art space remains open to you with no § 315(e)(2) estoppel currently running against anyone.

Per-proceeding detail

No proceedings to report

Because the structured feed is empty and my searches found nothing on the public dockets, there is no proceeding number, petitioner, panel, ground, institution decision, FWD, settlement, or appeal to describe. I am deliberately not populating the template with placeholders — any IPR20XX-XXXXX I typed here would be fabricated.

For completeness, the surrounding facts that make the null result interpretable:

• Patent: US 11,131,552 B2, "Map generation system." Filed 2019-12-06; granted 2021-09-28; priority to JP 2018-231019 (2018-12-10). Anticipated expiration 2039-12-06.
• Owner: The University of Tokyo and Toyota Jidosha Kabushiki Kaisha (corrective assignment recorded 2021-08-12). This is an OEM/university research asset, not a monetization vehicle. The original assignees and inventors (Tohriyama, Kamata, Ito, Nakamura) are the patent owner's own personnel — there is no assignment-to-NPE trail on the face of the record.
• Claims: 4 total. Claim 1 is the sole independent claim (server-side method of transferring reference curved-road constituent points to a measurement-target curved road as virtual constituent points, deriving virtual-trajectory curvature, and associating it with measurement-target constituent points). Claims 2–4 are dependent. Claim 4 recites the concentric-arc/center-point relationship between the partial virtual trajectory and the corresponding reference curved road arc.
• Forward citations: Only two patent families cite this patent (CN112050821B, CN113932794B — both Chinese lane-line/semantic-map updaters), which is a low-citation signal consistent with a patent that has not been litigated or asserted at scale.

Strategic summary

Claim status. All of claims 1, 2, 3, and 4 of US 11,131,552 are UNTESTED. Not canceled, not confirmed. There is no IPR certificate of cancellation, no certificate of correction narrowing a claim, and no district-court judgment of invalidity reported against this patent. The full claim set that issued on 2021-09-28 is the full claim set in force today. Claim 1 is the only independent claim, so any infringement case rises or falls on it plus its dependents; if you invalidate claim 1, claims 2–4 fall with it (they add limitations but depend from claim 1).

Estoppel landscape. There is no § 315(e)(2) estoppel in effect against any party, because § 315(e) is triggered only by a final written decision under § 318(a). Consequently:

• Every prior-art ground is available to you. No petitioner has locked up any reference.
• You face no § 315(b) one-year clock unless and until you are served with a complaint alleging infringement — and note that this being a patent held by Toyota and a university, the practical assertion risk is asymmetric (defensive cross-licensing context) rather than troll-driven.
• Conversely, you also cannot borrow anyone else's win. There is no FWD to cite, no claim-construction ruling from the Board under the Phillips standard, and no PTAB claim-construction record to leverage in a district court Markman.

Pattern signals. None of the classic patterns are present. No repeat petitioner (no petitioner at all). No patent-owner appellate aggressiveness (no CAFC appeal docketed). No defensive aggregator — Unified Patents, RPX, or similar does not appear anywhere in the chain; this patent has not been targeted by a crowd-funded validity challenge, which is itself notable given the patent sits in the hot autonomous-driving / HD-map space where Unified has been active.

Why the absence is plausible rather than anomalous: IPRs are filed in response to assertion, and this patent shows no assertion history. The AIA trial bar follows the money. A low-citation, OEM-and-university-held map-generation patent that has never appeared in a complaint will not attract a petition — the § 315(b) clock never starts, so prospective challengers rationally wait. That means the null PTAB record here is a low-signal absence, not the strong "nobody could kill this patent" signal you'd get from a litigated patent with zero IPRs.

One substantive observation you can use: the specification's own background art cites JP 2013-168016 A (Toyota) — the "straight road estimation from probe trajectories" reference — and the prosecution citations include Denso, Navteq/HERE, BMW, Continental, and Volkswagen art on lane-course determination, geometry creation for ADAS, and curve modeling. That is a dense obviousness landscape in exactly the claim's field, which is what makes this an attractive IPR candidate if it is ever asserted.

Recommended next steps

If you are a defendant being asserted against:

1. Do not cite any IPR. There is no FWD, no IPR certificate, and no Board disposition to link. The correct statement to your client and to the court is: "US 11,131,552 has never been before the PTAB; all four claims are untested and presumptively valid."
2. Treat the claim set as fully live. Because claim 1 is the only independent claim, target it. Claims 2–4 add (a) the lane-mark-crossing + curvature-center-intersection limitation, (b) the camera-based center-line estimation and constituent-point reset, and (c) the concentric-arc center-point limitation. These are narrowing; the broadest reach is claim 1.
3. Consider filing your own IPR — and note the timing benefit: if you have not been served with an infringement complaint naming this patent, no § 315(b) one-year bar is running, so you can prepare a petition on your own schedule. The procedural window is open and no one has taken the best art. Link your petition to the public record at PTAB E2E / PTAB Center once filed.
4. Erroneous-judgment exposure is nil — there is no Board decision to challenge, so no § 319 appeal and nothing on the Federal Circuit docket or CourtListener to monitor.
5. Re-verify before you rely on this. The ODP ingest can lag. Check USPTO Patent Center for US 11,131,552 and the PTAB E2E docket directly on the day you file anything. Family note: the JP sibling is JP 7060496B2, so any opposition/nullity attack there is a separate track and does not affect US claim scope.

If you are the patent owner: the record is clean but bare. You have no IPR survivorship to advertise, and the low forward-citation count plus the crowded ADAS/map prior-art field means the first real assertion of this patent is likely to draw a petition. Any licensing or enforcement posture should assume claim 1 will be tested.

Constraint acknowledgments: No proceeding numbers are reported because none exist on the record as of 2026-09-29. No claim-level disposition is characterized because no FWD exists. All dates are YYYY-MM-DD. This is an analyst's read of public records, not legal advice — confirm the PTAB docket and any newly-filed petitions before relying on the null result.

I'll research the assignment record for US 11,131,552 using the sources specified.

The initial broad searches were noisy. Let me search more precisely, including the reel/frame reference I can extract from the Google Patents legal events.

Let me try more targeted sources for the assignment record details.

Assignment & Ownership Analysis — US Patent No. 11,131,552 B2 ("Map generation system")

Sourcing caveat up front. I was not able to load records directly from USPTO Assignment Center (assignmentcenter.uspto.gov) within this session — my queries to it and to the mirror/indexed copies returned unrelated material or failed. The assignment facts below are reconstructed from the Google Patents legal-events record reproduced in the authoritative full text supplied for this analysis, which mirrors the USPTO assignment record. The reel/frame 051202/0424 is itself a second-hand citation — it appears inside a corrective-assignment entry that references the earlier recording. Treat the reel/frame as high confidence but not directly verified. Verify at: https://assignmentcenter.uspto.gov/ (search patent number 11131552) — I flag this explicitly rather than dressing up an unverified lookup as a direct read.

Inventors

Pattern notes (nothing alarming):

• This is a classic industry–academia joint filing: two inventors from the corporate co-assignee and two from the university co-assignee. That is a normal sponsored-research arrangement, not the "all inventors departed within 12 months" precursor to a fire-sale.
• No evidence of inventor departure, job change, or inventor-to-entity re-assignment was found. Critically, the inventors are listed as the assignors on both the original and the corrective assignment — i.e., they assigned their rights to the two institutions and are not residual holders of any interest.

Original assignee

Original assignee(s) of record: Toyota Jidosha Kabushiki Kaisha (listed on Google Patents as "Toyota Motor Corp") and The University of Tokyo (listed as "University of Tokyo NUC" / "TOYOTA JIDOSHA KABUSHIKI KAISHA, THE UNIVERSITY OF TOKYO" on the assignment events). Google Patents lists the Original Assignee and the Current Assignee as the same two entities — a strong affirmative indicator that no ownership transfer has occurred since issuance.

• Line of business / product embodiment: Toyota is the world's largest automaker and ships vehicles with ADAS and automated-driving features that depend on high-definition road/curve map data — the exact subject matter of the claims ("generate a map … for autonomously controlling an autonomous vehicle," claim 1). The University of Tokyo is a national research university that operates an autonomous-driving research program. Claims are directed to a map-generation server, not a consumer product, but the patented pipeline plausibly feeds Toyota's HD-map stack.
• Current status: Both operating. Neither is acquired, dissolved, or in bankruptcy. There is no corporate reorg, merger, or name-change conveyance in this chain other than the typo correction described below.

Assignment timeline

1. 2019-12-06 (executed) / 2019-12-06 (filing-date listing) — Reel 051202 / 0424

• Conveyance: ASSIGNMENT OF ASSIGNORS INTEREST (see document for details)
• Assignor: NAKAMURA, Satoshi; KAMATA, Minoru; ITO, Takuma; TOHRIYAMA, Kyoichi (all four inventors)
• Assignee: TOYOTA JIDOSHA KABUSHIK KAISHA (sic — "KABUSHIK" missing the final "I") and THE UNIVERSITY OF TOKYO
• Correspondent: Not captured / not verified. I could not retrieve the attorney or firm of record for this recording. (Note only, not an assignment-record finding: Toyota's US prosecution/trademark correspondence of record is typically Oblon, McClelland, Maier & Neustadt, L.L.P., 1940 Duke Street, Alexandria, VA 22314, per unrelated TTAB/patent records surfaced in search — but I have no confirmation that Oblon filed this assignment, so I do not assert it.)
• Context: Initial inventor assignment — the four inventors assigning their rights to the two co-owning institutions at filing. This is the standard, benign first link.

2. 2021-08-12 (recorded) — Reel/frame NOT CAPTURED

• Conveyance: CORRECTIVE ASSIGNMENT ("to correct the name of the first assignee previously recorded on reel 051202 frame 0424")
• Assignor: NAKAMURA, Satoshi; KAMATA, Minoru; ITO, Takuma; TOHRIYAMA, Kyoichi
• Assignee: TOYOTA JIDOSHA KABUSHIKI KAISHA (now spelled correctly) and THE UNIVERSITY OF TOKYO
• Correspondent: Not captured. If identical to the original recording, this would be a relevant repeat-correspondent datapoint — but I cannot confirm it.
• Context: Change of name / clerical correction only — it fixes the assignee's corporate name ("KABUSHIK KAISHA" → "KABUSHIKI KAISHA"). No change in ownership, no change in parties, no new consideration. Recorded roughly six weeks before grant (2021-09-28).

No other recorded assignments exist. There is no transfer to any "IP / Holdings / Licensing / Ventures" entity, no security agreement, no license recordation, and no release.

Timeline diagram

timeline
    title Ownership of US 11131552
    2018 : JP priority application filed
    2019 : US application filed
         : Inventors assign to Toyota and UTokyo
    2021 : Corrective assignment fixes assignee name
         : US patent granted

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. No assignment to any LLC or entity bearing an IP/Holdings/Licensing/Ventures suffix exists. Google Patents lists original assignee = current assignee (Toyota + University of Tokyo), meaning rights never left the operating company and the university. Reel 051202/0424 runs toward Toyota/UTokyo, not away from them.

2. Known asserter in the chain — NOT PRESENT. Neither Toyota, The University of Tokyo, nor any of the four inventors appears on any public NPE/asserter list (Acacia, Marathon, IV, IPNav, Wi-LAN/Conversant, Vringo, Pendrell, Round Rock, Spangenberg entities, etc.). Neither assignee is an NPE by construction — one is a global automaker, the other a national university. (Limitation: I did not complete a direct Unified Patents / RPX directory query.)

3. Repeat correspondent across the chain — UNCLEAR / NOT ASSESSABLE. I could not retrieve the correspondent of record for either recording (reel 051202/0424 or the 2021-08-12 correction). Because two recordings exist for a single family, a recurring correspondent would be the natural expectation — but with zero correspondent data captured I cannot mark this present, and I will not infer it. This is the one signal a direct Assignment Center lookup could still flip.

4. Cascading transfers — NOT PRESENT. Only one substantive assignment plus one clerical correction, spanning ~20 months. No chained LLCs, no shared-principal entities, no <24-month serial transfers.

5. Pre-litigation transfer — NOT PRESENT. Per the prior Litigation section, no infringement suit involving this patent is known. There is therefore no suit to time a transfer against, and no post-issuance transfer at all.

6. Bankruptcy fire-sale — NOT PRESENT. Neither assignee has filed for bankruptcy; no Chapter 7/11 sale of this patent appears in the record.

7. Privateering — NOT PRESENT. No operating company → NPE transfer; both co-owners retain their interests jointly.

8. Defensive aggregator — NOT PRESENT. The chain terminates at the two original co-owners, not at RPX, AST, LOT, Unified, or OIN. The patent is held, not neutralized.

Verdict

Insufficient data — with the affirmative caveat that every observable indicator runs against an NPE pattern, not toward one.

Justification. The record contains only the original inventor→assignee assignment at reel 051202/0424 (executed/listed 2019-12-06) and a clerical corrective assignment recorded 2021-08-12 that fixes a misspelling in the assignee's corporate name ("KABUSHIK KAISHA" → "KABUSHIKI KAISHA") with no change in parties or ownership. Google Patents lists the original and current assignee as the identical pair (Toyota Jidosha Kabushiki Kaisha and The University of Tokyo), and the prior Litigation section found no assertion activity, so there is nothing to trigger a pre-litigation-transfer or privateering finding. Under the strict rubric, "only the original assignment" is the Insufficient data case; but this is the benign flavor of insufficient data — a jointly-held industry/university patent that has never been separated from its owners. The one unresolved item is the correspondent of record on both recordings (signal 3), which a direct Assignment Center by-patent-number lookup at https://assignmentcenter.uspto.gov/ would resolve.

Flagged limitations (per operating rules, stated rather than papered over):

• Reel 051202/0424 is corroborated only through the corrective-assignment entry's cross-reference, not a direct USPTO read.
• The reel/frame of the 2021-08-12 corrective assignment and the correspondent(s) of record were not retrieved.
• My independent searches for this patent's USPTO assignment records returned irrelevant material; I did not fabricate any reel, frame, attorney name, or date to fill the gaps.

I'll verify the patent record and then work through its cited prior art. Let me run targeted searches.

Let me pull details on the most technically relevant cited references.

Prior-Art Analysis — US 11,131,552 B2 ("Map generation system")

Scope and verification note

Building on the bibliographic and claim sections already produced, I re-confirmed the record and worked through the citation list. Two process notes before the analysis:

• Source of the citation list. The USPTO-derived citation data is the "Citations (22)" table rendered on the Google Patents page for US 11,131,552. I treated that list as the authoritative set of references cited against this patent. A second rendered table ("Patent Citations (23)") appears in the page but is truncated in the source text — it breaks off at US 2013/0080019 A1 ("Vehicle behavior…"). I therefore cannot identify the 23rd entry, and I flag that as a known gap rather than guessing. The two rendered tables overlap heavily and appear to be the same body of art sorted differently.
• Verification limits. Full text was successfully retrieved and verified for a subset of references (US 9,448,074 B2; US 2011/0218724 A1 / US 8,583,366 B2; US 2015/0316386 A1; JP 2013-168016 A). For the remaining references I was unable to pull full text within this session, so their descriptions rest on the titles and dates in the patent's own citation list and are marked as unverified. Do not treat those one-line descriptions as a substitute for reading the documents.
• The rendered table does not preserve the examiner-cited (*) vs. third-party-cited (†) markers, so I cannot tell you from this data which references the examiner relied on versus which were submitted by a third party. That distinction matters for a § 102/§ 103 posture and should be confirmed against the file wrapper.

Critical date: the effective filing date / priority date is 2018-12-10 (JP 2018-231019). The application was filed 2019-12-06, after the AIA first-inventor-to-file date, so AIA § 102/§ 103 governs.

Framework used for the § 102 assessment

For each reference I state which subsection makes it available and which claim(s) it could potentially anticipate:

• § 102(a)(1) — the reference published (or was otherwise publicly available) before 2018-12-10.
• § 102(a)(2) — the reference is a U.S. patent or published U.S. application effectively filed before 2018-12-10 (this sweeps in references whose publication date post-dates the priority date but whose filing date does not).

A frank up-front conclusion, so it frames the table: on the face of the citation list and the material I could verify, no single cited reference discloses the full combination of claim 1 — in particular the transfer of reference-road constituent points across the lane mark onto an adjacent road as virtual constituent points, followed by setting a virtual trajectory from those virtual points plus the source shape information, and associating the resulting curvature information with target-road constituent points derived from travel distance. These references read as § 103 combination art, not clean § 102 art. Two structural facts drive that: (i) every one of the cited references appears to operate on a single road/lane from its own data, whereas claim 1's entire point is cross-road reuse; and (ii) none of them, so far as I can tell, construct the curvature center as the intersection of the lines joining each virtual point to its corresponding reference point (claim 2) — which is the geometric heart of the disclosure.

The cited references, one by one

Tier 1 — Closest to the claim's subject matter (verified content)

US 9,448,074 B2 — Denso Corp. — "Curve modeling device, curve modeling method, and vehicular navigation device" — JP priority 2013-01-23; granted 2016-09-20. § 102(a)(1).
Verified content: computes a curvature at each sampling point on a route; a correction unit corrects the curvature under the condition that an orientation difference defined by the calculated curvatures stays constant; the route is approximated by straight-line / arc / clothoid intervals; a node-information generation unit builds a shape model and emits node information.
§ 102 assessment: Does not anticipate claim 1. It is missing the cross-road transfer, the virtual constituent points, and the "second detailed curve information" association step. Closest to claim 1's element (f) (deriving shape data for spaced points along a route) but is single-road. Best characterized as § 103 art against the "curvature-based shape modeling" concept.

US 2011/0218724 A1 (granted as US 8,583,366 B2) — Denso Corp. — "Road shape learning apparatus" — priority 2010-03-04; published 2011-09-08. § 102(a)(1).
Verified content: amends an entrance coordinate, middle coordinate and exit coordinate of a traveled curve using amendment values that depend on travel tendency, then computes the radius of the circular arc passing through the three amended coordinates and designates it as the curve's curvature radius; stores that radius as learned data associated with the map data so the map is "improved to be more detailed so as to be suited to the actual shape of the curve." A second embodiment generates control data for vehicle behavior/velocity from the learned data.
§ 102 assessment: Not an anticipation of claim 1, but this is the single closest reference to the patent's core idea of "deepening" a map by associating a learned curvature radius with map points — the specification's own description of the prior "deepening" method. Missing: transfer to an adjacent road; virtual constituent points; virtual trajectory. Relevant to claim 1 element (f). Good § 103 art, especially when combined with a cross-lane geometry reference.

JP 2013-168016 A — Toyota Motor Corp. (inventor Kashiwai Tadahiro) — "Travel lane recognition device" — priority 2012-02-15; published 2013-08-29. § 102(a)(1). (This is also the reference the patent's own Background section discusses.)
Verified content: acquires trajectory information for a plurality of vehicles that have traveled a road section and estimates the lane of that road section from that trajectory information — expressly so that a lane can be estimated in real time without the subject vehicle having first driven the road.
§ 102 assessment: Not an anticipation. It uses fleet trajectory data and is directed at lane recognition, not at generating detailed curve information by transferring an adjacent road's shape data. It is closest to claim 1's reliance on "travel information of a vehicle traveling on the measurement target curved road" — and it is the reference the examiner/specification singled out, so it is the natural starting point for a § 103 obviousness attack on the "use of travel information" element.

US 2015/0316386 A1 — Toyota Motor Engineering & Manufacturing North America (inventor Delp) — "Detailed map format for autonomous driving" — priority 2014-04-30; published 2015-11-05. § 102(a)(1).
Verified content: a computer-readable map format with lane segments formed of waypoints and border segments formed of borderpoints; a lane width of a lane segment can be determined by measuring the distance between two border segments on opposite sides; border type/color drive driving rules; intended for autonomous-vehicle control and localization.
§ 102 assessment: Not an anticipation of any of claims 1–4. It supplies the "waypoint/width" vocabulary (relevant to claim 1's "constituent points" and the lane-width W in the transfer equations) and touches claim 1's "for autonomously controlling an autonomous vehicle" field of use, but it discloses no curvature transfer between roads.

Tier 2 — Same field (map building / trajectory / lane geometry); titles and dates from the citation list, full text not verified this session

Tier 3 — Older/broader background art (titles/dates from citation list)

Family-cited references (cited in the family, not necessarily against this patent)

Ranking: the most relevant prior art

If you need a short list for a validity/§ 103 analysis, this is my ordering:

1. US 2011/0218724 A1 (Denso, road shape learning apparatus / US 8,583,366 B2) — teaches the learned-curvature-radius-associated-with-map-data "deepening" concept that the patent's specification adopts as its own premise. Primary art for element (f).
2. US 9,448,074 B2 (Denso, curve modeling) — teaches deriving arc/clothoid shape models and node information from sampled curvature along a route. Primary art for the shape-modeling and curvature-radius aspects of claim 1.
3. JP 2013-168016 A (Toyota, travel lane recognition device) — teaches using travel trajectory information of vehicles on a road section to estimate a lane. Primary art for the "travel information" element; also the reference the patent itself distinguishes in its Background.
4. US 2015/0316386 A1 (Toyota, detailed map format for autonomous driving) — teaches waypoint/constituent-point lane representation, lane width measured between border segments, and autonomous-vehicle use. Primary art for the waypoint/lane-width vocabulary and the autonomy field of use.
5. US 2020/0122721 A1 (Baidu, two-step reference line smoothing) — closest cited art to claim 1's "set the virtual trajectory" and to claim 4's arc geometry, and it is § 102(a)(2) art (effectively filed 2018-10-23).

What is not in any cited reference, on my reading: (i) transferring reference-road constituent points across the lane mark onto an adjacent target road; (ii) the claim 2 construction of the curvature center as the intersection of the lines P_iQ_i and P_{i+1}Q_{i+1}; (iii) the claim 3 camera-based center-line estimation and resetting of constituent points; and (iv) the claim 4 concentric-arc limitation. Those four items are the patent's apparent point of novelty, and the cited art appears to be combination material directed at the individual elements rather than at that combination.

Caveats you should carry forward

• This is an analyst assessment, not a legal opinion or a claim chart. The "potentially anticipates" assignments for Tier 2 and Tier 3 references are provisional and rest on titles/dates, not on read text. Each would need to be read in full against claim 1's five elements before any § 102 position is asserted.
• The 23rd citation in the second rendered table is missing from the source text, so my list may omit one examiner- or third-party-cited reference.
• Examiner-cited vs. third-party-cited status is not recoverable from the rendered data; it should be confirmed from the file wrapper, as it changes the weight of the prior-art record.
• No litigation, PTAB, ITC, or Federal Circuit activity was found for this patent (per the earlier sections), so there is no adjudicated prior-art record or claim construction to draw on.
• No auto-correction applied. Patent numbers, application numbers, publication numbers, reference numbers and dates appear exactly as the records render them (including the two recorded spellings of the Toyota assignee name discussed earlier).

I'll ground this analysis by pulling what public sources actually say about the key cited references, then map them to the claim elements. I'll also flag a date inconsistency I noticed between the prior sections and this task.

Obviousness Analysis — US 11,131,552 B2 under 35 U.S.C. § 103

Analyst: Senior US patent analyst
Task date as given: 2026-04-26
Source of record: The prior-art section of the Google Patents page for US 11,131,552 B2 (Citations (22) / Patent Citations (23) / Family Cites Families (3)), supplemented by live retrieval of reference text where available.

0. Two threshold flags before the analysis

(a) Date inconsistency across the file (must be reconciled). The earlier sections of this report are dated 2026-09-29; this task states 2026-04-26. I am not silently harmonizing them. The patent record itself, the PTAB null result, and all bibliographic data are date-stable, so nothing substantive turns on the discrepancy — but if a single "as-of" date must be certified, resolve it before filing.

(b) Data-quality inconsistency inside the source page. The page renders both "Citations (22)" and "Patent Citations (23)." I did not auto-correct either number. The 22-row table lists exactly 22 distinct publication numbers; the "23" likely reflects a duplicate or an unrendered row. Separately, every row carries the * marker, which the page legend defines as "Cited by examiner." If that legend is accurate, this art is of record and was not used to reject. That is a material fact for any § 103 challenge and I address it in § IX.

I. Legal framework applied

I apply the Graham v. John Deere Co., 383 U.S. 1 (1966), factors as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007):

1. Scope and content of the prior art;
2. Differences between the prior art and the claims;
3. Level of ordinary skill in the art; and
4. Objective indicia (secondary considerations).

Combination efficacy is judged under the KSR rationales (MPEP 2143): known technique to improve similar devices; simple substitution of a known element; use of a known technique to improve a similar device in the same way; application of a known technique to a known device ready for improvement; "obvious to try"; and market/design incentives. References need not be physically combinable — only logically combinable (In re Keller, 642 F.2d 413 (CCPA 1981)). The claimed result need only be predictable, not proved in the art (In re McLaughlin).

II. Level of ordinary skill (POSITA)

A POSITA here is a person with a bachelor's degree in electrical/mechanical engineering, computer science, or geomatics, plus roughly 2–4 years of experience in digital road-map creation, vehicle positioning, or ADAS/automated-driving path planning — or equivalent. This is a predictable, well-developed, engineering intensive art (map geometry; curvature mathematics; probe-data aggregation), which matters because KSR accords strong weight to the "predictable variations" rationale in such arts.

III. Prior-art status of the references on the face of the patent

The critical date is 2018-12-10 (JP 2018-231019). A publication published after that date is not available under § 102(a)(1) as a printed publication; it is available only under § 102(a)(2) (secret prior art), and only against subject matter it was "effectively filed" for. This distinction is frequently botched, so I have segregated the list. (Dates below are the page's "Priority date" column, which the page itself labels an assumption.)

III-A. Available as § 102(a)(1) art (published before 2018-12-10)

III-B. Available only as § 102(a)(2) art (published 2019–2020; pre-critical-date priority)

US 2019/0025063 A1 (VW); US 2019/0113925 A1 (Mando); US 2019/0196472 A1 (Continental — analyzing driving trajectories for a route section); US 2019/0360819 A1 (HERE); US 2019/0391594 A1 (Nissan); US 2020/0122721 A1 (Baidu); US 2020/0124424 A1 (Showa); US 2020/0139959 A1 (Zoox); US 2020/0141738 A1 (HERE); US 2020/0166364 A1 (Nissan); US 2020/0225044 A1 (Toyota — CAUTION: common-ownership exclusion, see below); US 2020/0348146 A1 (Denso).

Common-ownership caution (§ 103(c) / § 102(b)(2)(C)). Any reference qualifying only under § 102(a)(2) that was, at the time the invention was made, owned by or under an obligation of assignment to Toyota Jidosha Kabushiki Kaisha or The University of Tokyo is disqualified as prior art. US 2020/0225044 A1 (Toyota) falls in that bucket and I would not build a ground on it. US 2020/0166364 A1 (Nissan), US 2020/0122721 A1 (Baidu), US 2020/0141738 A1 (HERE), US 2020/0139959 A1 (Zoox), US 2020/0124424 A1 (Showa), and US 2019/0196472 A1 (Continental) are not commonly owned and remain available.

III-C. Family-cited JP art (available if translated/verified)

JP 2014-197351 A (Aisin AW, 2014-10-16); JP 2015-075423 A (Denso, 2015-04-20); JP 6564618 B2 (Aisin AW, 2019-08-21 grant). The first two are § 102(a)(1) art if properly published and translated. I have not retrieved their texts and do not rely on them below beyond flagging them as an unexplored source.

IV. Claim 1 — element-by-element mapping

Structural observation carried forward from the earlier sections (not repeated here): claim 1's final clause is phrased as an intended use. Under In re Sinex, 309 F.2d 488 (CCPA 1962), and the § 2114 line of authority, a bare statement of intended use adds little; and even if given full weight, Toyota's own US 2015/0316386 was published 2015-11-05 and expressly targets autonomous driving. So this clause, likely added during prosecution to secure allowance, does not create distance from the art.

V. Specific § 103 combinations

Ground 1 (strongest for claim 1): Navteq '716 + Denso '074 + Toyota '386

• Navteq '716 supplies the transfer: geometry for a first road (motorway nodes/shape points) is offset to create curve-fit points for an adjacent road (ramp spline), stitched into a single spline, with curvature/slope/heading computed from the resulting curves.
• Denso '074 supplies the curvature engine: compute curvature at each sampling point of a path, correct it, approximate the path as straight-line / arc / clothoid intervals, and generate node information forming a shape model — the exact "virtual trajectory curvature information" mechanics of claim 1. Denso '074 also supplies the motivation frame: it states existing navigation maps "are not accurate enough to provide a driving control yet."
• Toyota '386 supplies the "detailed map… for autonomously controlling an autonomous vehicle" element and the waypoint/lane-width data model (lane segments of waypoints; border segments; lane width measured between border segments; lane centers determined from lane borders).

Motivation: all three sit in the same field of endeavor (digital road-map geometry for driver assistance/automation) and address the same problem — producing ADAS-grade curve geometry cheaply. Navteq's offset step exists precisely because re-deriving full spline geometry for every adjacent roadway from independent surveys is wasteful; Denso '074's curvature-and-arc modeling is the known way to convert sampled positions into usable curve parameters; Toyota '386 is the known output format. Combining them is the "known technique to improve a similar device in the same way" rationale.

Ground 2 (strongest motivation showing): Denso '724 + JP 2013-168016 + Toyota '386

This ground is built to destroy the applicant's stated advantage, and I regard it as the most dangerous to the patent.

• Denso '724 expressly computes a curvature radius as the radius of a circular arc passing through entrance/middle/exit coordinates, and stores it as learned data associated with the map data — i.e., "deepening." It further states that a curvature radius obtained from a travel track in one lane differs from the curvature radius pertinent to the center of the road width, and it defines its correction values "as a distance up to the center line" (a value "corresponding to a distance from a center in a road width direction"). Its stated effects are: "the position information on the center line is memorized in the learned data; the amended shape of the road can be acquired by traveling in one traffic direction without need to travel in two opposite traffic directions… the work required for the learning can be simplified" and the stored data and storage capacity are reduced.

  That is, by 2011 Denso had expressly taught acquiring the road shape for a lane you did not drive, by offsetting across the road-width direction, to reduce the number of required passes.

• JP 2013-168016 A (Toyota; the applicant's own background reference) teaches the complementary half: it recognizes the defect of host-vehicle-only accumulation — "since the system… can accumulate a travel track only after the host vehicle travels the road whose lane is to be defined, it can only estimate lanes of roads the host vehicle has traveled" — and solves it by acquiring trajectory information of a plurality of vehicles on a road section and estimating the lane from that.

• Toyota '386 supplies the detailed-map/autonomous-control output.

Motivation: KSR attaches near-decisive weight where the prior art articulates the very advantage the applicant asserts. Here the applicant's stated benefit is "the number of travels of the vehicle V on the measurement target curved road necessary to make the detail curve information… have a certain accuracy or more can be reduced." Denso '724 states the same benefit ("without need to travel in two opposite traffic directions… the work required for the learning can be simplified"), and JP 2013-168016 states the same problem (coverage of roads not personally traveled). Under KSR, an advantage taught or suggested by the art is not a patentable contribution.

Ground 3 (server-side, probe-data architecture): Continental '472 + Denso '074 + Navteq '716

• Continental '472 discloses a vehicle-external system with interfaces, a database, and data-processing modules that receive first trajectories (series of waypoints) from vehicles traversing a route section, cluster them, and generate/store second trajectories for that section, then serve them to vehicles "for controlling an autonomously driving vehicle." Fig. 3 shows trajectories along a curve where the typical path "does not always run in the middle of the lane, but rather runs at different distances between the outer edge of the roadway and the center line" — i.e., an explicit recognition of lateral offset between a driven path and lane geometry, and of analyzing the driving envelopes of adjacent lanes.
• Denso '074 supplies the curvature/arc modeling; Navteq '716 supplies the cross-lane offset transfer.

Motivation: Continental's own stated object is generating section-specific trajectories where vehicle sensors are inadequate — the same "fill in the un-traveled road" problem. Combining probe aggregation with a map-geometry offset engine and a curvature model is the combination of known elements each performing its known function.

VI. Dependent claims

Claim 2 — lane-mark crossing; curvature center by intersecting two radial lines; curvature radius as center-to-point distance

• Lane-mark crossing: Navteq '716's ramp offset crosses exactly the divide between adjacent roadways; Denso '724 and Toyota '386 both operate relative to lane boundaries/border segments.
• Curvature center via intersection of the line through P_i–Q_i and the line through P_{i+1}–Q_{i+1}: this is the standard center-finding construction. It is disclosed in substance by Denso '074, which recites the earlier art (JP H9-185322) computing "a curvature at each node… by calculating an approximate expression of a circle that passes through the three neighboring node points" — the perpendicular-bisector/radial-line intersection being the identical geometry. Denso '724 goes further and expressly computes straight lines P0–A0, P0–B0, P0–C0 through a common center point P0 and moves points along those lines. A POSITA seeking the center of an arc through two transferred points would use the lines joining each transferred point to its source point; that is elementary geometry, not invention (KSR, "a finite number of identified, predictable solutions").
• Curvature radius = distance from center to the adjacent points: Denso '724 ("computes a radius of the circular arc… designate[s] the computed radius as a curvature radius").
• Conclusion: claim 2 adds a geometry recipe that is either disclosed or a predictable mathematical consequence of the claim 1 transfer. Obvious.

Claim 3 — camera-based center-line estimation and resetting constituent points on the center line

• Toyota '386: lane borders are captured "using the above described LIDAR system and/or cameras," and "lane centers can be determined based on the lane borders," with lane width determined by measuring between border segments.
• Continental '472: imaging sensors determine position within a lane "with reference to… features of the roadway or lane that are fundamentally invariable or change only slowly over time, such as roadway markings."
• Denso '724: the amendment is directed to the center in the road width direction, using values representing the "distance up to the center line" — a direct precursor to "reset the measurement target curved road constituent points on the center line."
• Caveat / genuine weakness: none of the above snaps a map constituent point onto a camera-estimated center line. The camera-to-map-point reset is the one step where I do not have a clean single-reference hit. Nonetheless, lane-mark detection and lane-center estimation from camera imagery is one of the most notoriously well-known techniques in this art (the patent itself claims the benefit of ADAS features including LKA, and the examiner-cited JP 6564618 B2 — "road shape detection system" — sits in this space but I could not retrieve its text). I would characterize claim 3 as obvious over Toyota '386 + Continental '472 + Denso '724, but with moderately lower confidence than claims 1 and 2, and recommend targeted retrieval of JP 6564618 B2 and US 2016/0039413 A1 (BMW) before finalizing.

Claim 4 — the partial virtual trajectory arc and the corresponding reference-road arc share a center point

This is the claim I consider most vulnerable to a "pure geometry / predictable result" attack, for a reason the specification itself concedes without noticing:

The offset curve of a circular arc is itself a circular arc sharing the same center, with radius r ± W. The patent's own transfer equation uses a single offset W = (W₁ + W₂)/2 applied along the local normal (X_i = x_i + W·cos(θ_i − π/2)). Therefore, for any arc-shaped road, concentricity is the mathematical consequence of the claimed offset, not a separate inventive step. KSR treats "a predictable variation" and results that are "the product … of ordinary mathematical calculation" as obvious. Denso '724 makes this explicit structurally, moving points along lines through a common center P0 and recomputing the radius from the moved points.

Conclusion: claim 4 is a dependent claim, and if claim 1 is invalid, claim 4 falls with it. Even standing alone, claim 4 reads on the inherent geometry of the claimed transfer. Obvious.

VII. Consolidated motivation to combine (KSR rationales)

No teaching away. Denso '724's recognition that travel tracks contain "out-in-out" bias argues for correction-and-reuse, not against it; JP 2013-168016's multi-probe averaging removes individual-driver bias. Neither reference disparages cross-lane reuse. In re Gurley / DePuy teaching-away requires a criticizable, exclusionary teaching, which is absent.

Reasonable expectation of success. High. The output is a map-attribute value used for vehicle control; Denso '074, Denso '724, Toyota '386, and Navteq '716 each already produce usable geometric parameters for vehicle control at claimed-equivalent fidelity.

VIII. Objective indicia

None located. The earlier sections of this report establish no litigation, no PTAB record, and only two forward-citing families (CN112050821B, CN113932794B) — i.e., no evidence of commercial success, copying, licensing, or industry praise tied by nexus to the claims. Conversely, the absence of any assertion history makes it unlikely that objective indicia will be developed in time to rescue the patent. If the patentee were to assert the "fewer measurement passes" benefit, nexus fails for the reason in § VII(b): the benefit was already stated in the prior art.

IX. Where the patent could survive — honest counter-analysis

The patentee's best arguments, and my assessment:

1. The specific use of already-deepened reference-road detailed curve information as the seed for a not-yet-deepened adjacent road. This is the closest thing to a novel emphasis in the claim set. But Denso '724 already teaches deriving the road-center shape across the road-width direction from a single lane's learned data, and its purpose is to avoid a second pass. Weight: weak.

2. Associating a single curvature radius with whichever target-road constituent point falls between two virtual constituent points, where the virtual points do not coincide with target-road points. This is the most defensible novel step; the specification flags it as the point of the design. Counter: assigning a parameter to the nearest indexed sample, or to all samples within an interval, is routine data-indexing. Weight: moderate but likely insufficient, especially with claim 4's concentricity being inherent.

3. The intentional-use clause. I do not expect it to carry patentable weight (§ IV), and it is met by Toyota's own 2015 art even if it does.

4. Prosecution history caution. Because the specification's own Summary ends claim 1 at "generate a map by acquiring the second detailed curve information" without the autonomous-control phrase, that clause appears to have been added during prosecution. Two consequences: (i) it signals the applicant and examiner viewed the art as close, which cuts against nonobviousness of the remainder; and (ii) it raises a § 112(a) written-description / new-matter question, which is an independent invalidity theory worth developing. The examiner-cited status of all 22 references (§ 0(b)) means the art of record was considered but the examinee record shows the claim was allowed anyway — so a successful challenge will most likely need either (a) a combination the examiner did not articulate, or (b) the geometry/inherency argument against claim 4 plus a broad reading of Navteq's offset teaching.

X. Verification gaps and confidence

• Full text retrieved: US 9,448,074 B2; US 2014/0249716 A1 (and US 8,762,046 B2 / EP 2 172 748); US 2015/0316386 A1 family (US 10,118,614 B2, US 9,921,585 B2); JP 2013-168016 A (machine translation); US 2011/0218724 A1; US 2019/0196472 A1 / US 11,226,622 B2; DE 10 2016 216 335 B4.
• Not retrieved (analysis based on title/assignee/date only, so treated as corroborative, not load-bearing): US 2010/0305850; US 2012/0095682; US 2013/0080019; US 2016/0039413 (BMW); US 2018/0217612; US 2019/0025063; US 2019/0113925; US 2019/0360819; US 2019/0391594; US 2020/0122721; US 2020/0124424; US 2020/0139959; US 2020/0141738; US 2020/0166364; US 2020/0225044; US 2020/0348146; and the three family-cited JP documents. My retrieval run was truncated by a step limit, so several of these may contain express teachings stronger than what I attributed to them. Every one of them should be pulled in full before a petition is drafted.
• No fabrication: every proceeding number, quote, and URL used above comes from the record or from a fetched source. Where I have characterized a reference by title only, I have said so.
• Not legal advice. These are an analyst's grounds-oriented readings, not adjudicated constructions. Confirm the PTAB docket (https://ptacts.uspto.gov/ptabweb), Patent Center (https://patentcenter.uspto.gov), and the Federal Circuit docket before relying on any of it.

Grounding URLs used:

• https://patents.google.com/patent/US9448074/en
• https://patents.google.com/patent/US20140249716
• https://www.freepatentsonline.com/y2014/0249716.html
• https://patents.google.com/patent/US20150316386A1 (family; see also https://portal.unifiedpatents.com/patents/patent/US-[10118614](/patent/10118614)-B2)
• https://patents.google.com/patent/JP2013168016A/en
• https://www.patentsencyclopedia.com/app/20110218724
• https://patents.justia.com/patent/[11226622](/patent/11226622) and https://uspto.report/patent/app/20190196472

XI. Bottom line

All four claims of US 11,131,552 are vulnerable to § 103 attack. The strongest single combination is Denso US 9,448,074 B2 + Navteq US 2014/0249716 A1 (US 8,762,046 B2) + Toyota US 2015/0316386 A1, with Denso US 2011/0218724 A1 + JP 2013-168016 A supplying the decisive motivation: the prior art already taught (i) offsetting one road's node/shape-point geometry onto an adjacent road, (ii) deriving curvature from sampled path points and modeling it as an arc, (iii) storing the derived curvature as learned data associated with the map, and (iv) doing all of it to avoid having to drive the other lane or direction — which is precisely the advantage the applicant asserts. Claim 4 is additionally exposed because concentricity is the inherent mathematical result of the claimed offset of a circular arc, and claim 1's autonomous-control tail is met by the applicant's own sister filing from 2014–2015.

Claims 2 and 4 I assess as obvious with high confidence. Claim 3 is obvious with moderate confidence, pending retrieval of JP 6564618 B2 and US 2016/0039413 A1. If a challenge is filed, the priority-art status filters in § III (especially the § 102(a)(2)-only set and the Toyota common-ownership exclusion) must be applied before any ground is pleaded.