US 4792018

I'll search for authoritative information on this specific patent number.

Let me check for any current litigation or PTAB/Court of Appeals activity.

Let me verify the patent's legal status and check for any recent proceedings tied to that exact number.

US Patent 4,792,018 — Analyst Summary

Searches run: Google Patents full text, uspto.report grant page, FreePatentsOnline citation records, PTAB/PTACTS petition database, and general web queries for the literal strings 4792018, US4792018A, and "4,792,018" combined with litigation/appeal terms. No litigation, IPR/PGR, or Federal Circuit docket activity involving this patent surfaced. See "Uncertainty notes" at the end.

1. Bibliographic data

Source: https://patents.google.com/patent/US4792018/en ; https://uspto.report/patent/grant/4792018

2. Abstract (verbatim)

"A system for processing articles selected for purchase and bearing distinct identification codes comprises, in one version: (a) a code reader for generating an output signal indicative of such article identification code; (b) a conveyor for receipt and transport of such article; (c) an entrance sentry for defining an inlet to a secured zone extending along a portion of the conveyor and for generating an output signal indicative of entry of the article into the secured zone and of a measurable characteristic of the article; and (d) a controller for selective movement of the conveyor in respective article acceptance and article rejection senses. The controller is operable in several respects, namely, for storage, for each of a plurality of such articles, of a signal indicative of a predetermined value of the measurable article characteristic correlated with such article identification code, for response to the code reader output signal for comparison of such stored signal with the output signal of the sentry, and for operation of the conveyor selectively in response to the results of such comparison."

3. Plain-language overview of the independent claims

Claim 1 — Sentry-based shape/characteristic check (system). A checkout system has (a) a code reader, (b) a conveyor, (c) an "entrance sentry" at the mouth of a secured tunnel that does two jobs with one sensor — it signals both that the article has entered the zone and a measurable physical characteristic of that article (e.g. height/shape via the stepped LED/photocell light curtain of Fig. 5), and (d) a controller that stores a target value of that characteristic keyed to the article's ID code, compares the sentry measurement to the stored target when the code is read, and runs the conveyor forward (accept) or in reverse (reject) based on the comparison.

Claim 9 — Dual code reading plus EAS check (system). Same general architecture, but the security logic is: (d) a second code reader inside the secured zone, with the controller comparing the first (outside) code read against the second (inside) code read to catch substitutions; and (f) an EAS (electronic article surveillance) detector determining whether the article in the zone is EAS-tagged, with the controller storing, per article ID code, whether that article should be tagged and moving the conveyor based on stored indication vs. actual detection.

Claim 11 — Self-compiling ("SETUP mode") system. The controller is not dependent on a pre-existing store database: it itself builds the reference data by processing the code reader output and the characteristic sensor output (see the Fig. 13 SETUP flowchart, storing weight and up to three alternative heights H1/H2/H3 per UPC), and then performs the comparison of measured vs. self-compiled stored value and controls the conveyor accordingly.

Claim 12 — EAS-only system. A system whose security basis is: code reader + conveyor + secured zone + EAS detector generating an output signal, with the controller storing an indication correlated with the article ID code of whether the article should be EAS-tagged, comparing the detector signal against that stored indication, and selectively accepting/rejecting via the conveyor.

Claim 14 — Method with dual code read + EAS. Steps: (a) read the identification code; (b) put the article in the secured zone and there (I) measure a characteristic and (II)(A) read the code again and (B) subject it to EAS examination; (c) reject and return the article to the purchaser on (I) failure of the measured characteristic to correspond to a stored predetermined value, or (II)(A) failure of the second-read code to correspond to the first-read code and (B) failure of the article to pass the EAS practice.

Claim 15 — Method with characteristic + EAS (no second code read). Same as claim 14 minus the re-read of the code: reject and return the article upon (I) mismatch of the measured characteristic with the stored predetermined value, or (II) failure of the article to pass the EAS practice. Claim 16 narrows the characteristic to article shape; claim 17 adds weight measurement as a further rejection criterion.

Drafting notes worth flagging (quoted literally from the granted text)

• Claim 1(b): "conveyor means for receipt and transport of such aricle" — typographical error in the printed claim.
• Claim 9(f): "whether or not such article should be EAC-tagged" — typographical error.
• Claim 10: "a predetermining value of said article characteristic"; Claim 13: "in response to he results".
• Claim 14, condition II is written conjunctively ("A. failure of correspondence … and B. failure of such article to pass such EAS practice"), which on its face requires both the code mismatch and the EAS failure before the mandatory rejection of that branch — a scope quirk compared with claim 9's system-level counterparts.

4. Technical context (from the specification)

• The patent is a CIP of the '913 application and improves on Otis US 2,919,851 and the Zellweger '570/'571 patents (US 3,681,570 / 3,681,571) by adding multiple independent fraud-rejection bases: article shape/height, weight, an in-zone second UPC read, and EAS tag presence/absence.
• Hardware described includes entry sentry 33 and exit sentry 34 (LED array 31/35 + photocell array 32/36), optional horizontal curtain 37 and beam 42 in the bagging area, weight sensor 43 under entry conveyor 12, CPU 44, shape comparator 114, weight comparator 118, memory 140, and an X-Y driven inside UPC reader 130 (Fig. 6).
• Rejection is effected by reversing the entry conveyor; Fig. 13 describes a "SETUP" mode in which the system learns target weight and up to three acceptable heights per UPC, with a "completed record" flag (step 272).
• The specification's own fraud example: a customer scans an inexpensive wine bottle's UPC and puts an expensive, similarly sized/weighted bottle on the conveyor — the EAS aspect (Fig. 9, steps 206–215) is the designed countermeasure.

5. Litigation / CAFC 2026 search result

• No 2026 (or other) Federal Circuit docket, IPR, PGR, or district-court action involving US 4,792,018 was found in the sources searched. This is consistent with the record showing anticipated expiration on 2005-12-20 and legal status "Expired – Lifetime," which would moot enforcement.
• Searches for the string "4792018" repeatedly return US 7,920,018 (Advanced Memory Technologies v. SK Hynix, E.D. Tex., IPR2025-01450) — a different patent and not to be conflated with 4,792,018. I excluded those results as instructed.

6. Uncertainty / verification notes

1. Negative search results are not proof of absence. I could not access the USPTO PatentCenter/PAIR record or a complete CAFC docket database directly; the "no litigation" finding rests on the search results I could retrieve and on the expired status. Treat it as high-confidence but not authoritative.
2. "Current assignee" on Google Patents (Uniquest Inc / IBM / NewCheck Corp) is an automated compilation of recorded assignments, not a legal conclusion; the recorded IBM assignment (2005-04-25) postdates most of the patent's commercial life and falls just before expiration.
3. The "Citations" list in the supplied full text is truncated mid-entry at US 3,832,530 (Westinghouse), so the complete prior-art citation set for this patent is not reproduced here.
4. All bibliographic values above are taken from the authoritative Google Patents full text supplied for patent number 4792018 and cross-checked against uspto.report and FreePatentsOnline citation records; no conflicting values were found.

I'll search for litigation involving US patent 4792018.

Let me search more specifically for litigation and enforcement history.

Let me try litigation-specific databases.

Findings on litigation involving U.S. Patent No. 4,792,018

Based on the searches I was able to run (open-web searches targeting patent litigation aggregators, PTAB/IPR pleadings that list related litigations, and patent-family records), I did not locate any litigation that specifically asserts U.S. Patent No. 4,792,018. I want to be candid rather than fill in the blanks: I will not invent case numbers, courts, or dates.

What I could verify about the patent itself (from the document provided)

Because the patent's 17-year term from the 1988 grant (anticipated expiration 2005-12-20) has passed, any infringement action on the '018 patent itself would necessarily have been filed before that date, and the patent is now expired and unenforceable for prospective relief.

What my searches did and did not cover

• Did not find: Any entry in litigation-tracking content naming "4,792,018" as an asserted patent, and no IPR/PGR petition, complaint, or judgment referencing the '018 patent.
• Search results that surfaced instead were false positives — e.g., "JP 4792018 B2" (an unrelated Japanese patent) and a Brazilian corporate-registry entry containing the string "4792018." These are not U.S. 4,792,018 and I have excluded them.
• Coverage limitation (important): I do not have direct, reliable full-text access to PACER, Docket Navigator, or the Unified Patents litigation caselist through these tools, and the open-web results were dominated by patent-family citations (the '018 patent appears frequently as a cited reference in later patents, which is not litigation). My searches therefore cannot be treated as a complete litigation clearance.

Family/enforcement context (stated with appropriate caution)

The '018 patent shares a specification and priority with U.S. 4,676,343 (the '913 parent), both originally assigned to Checkrobot Inc., with the portfolio later passing to Uniquest, NewCheck Corporation / Productivity Solutions, Inc., and finally IBM in 2005. It is plausible that enforcement activity in the self-checkout space in the late 1980s–1990s was directed at these related patents, and it is common for such reporting to refer to the family generally rather than to the '018 patent by number. However, I have not confirmed any specific case asserting the '018 patent, and I am not going to attribute any particular case to it.

Recommended authoritative sources to confirm (I could not fully query these)

1. Unified Patents litigation caselist – https://portal.unifiedpatents.com/litigation/caselist (search by patent number)
2. CourtListener / RECAP – https://www.courtlistener.com (docket and opinion full-text search for "4792018" and "4,792,018")
3. PACER – civil docket search, including for the original assignee "Checkrobot Inc." and successors "NewCheck Corporation" and "Productivity Solutions, Inc." in districts such as D. Del., N.D. Ga., N.D. Ill., and S.D. Fla. (where store-automation and NCR/PSC/Optimal Robotics litigations clustered)
4. USPTO Patent Assignment Search and Patent Center – to trace chain of title (which establishes who had standing to sue)
5. Westlaw / Lexis / Docket Navigator – for historical 1988–2005 pleadings, which are often outside free databases

Bottom line

Based strictly on the searches performed and the sources available to me: No litigation specifically involving U.S. Patent No. 4,792,018 was identified. This should be read as "not found," not as an affirmative statement that none exists — the patent predates widely indexed free litigation databases, and I could not query PACER, the Unified Patents caselist, or Docket Navigator directly. If you can run the patent number through those databases (particularly Docket Navigator's "patent" field search and CourtListener), that would give a definitive answer.

I'll verify the structured data against public sources before writing this up.

Let me check the PTAB E2E system and any other public dockets directly.

Let me run a few final confirmation searches for CBM reviews and any related litigation history.

Proceedings overview

Zero AIA trial proceedings on file — the USPTO Open Data Portal returns no IPR, PGR, or CBM for US 4,792,018, and my independent web searches across PTAB-related sources surfaced no petition, institution decision, or Final Written Decision bearing this patent number (active: 0, claims invalidated: 0, claims sustained: 0, settled: 0, institution denied: 0). The bottom-line defensive posture is unusual but favorable: all 17 claims are UNTESTED at the Board, and the patent is expired — so the correct response to any demand letter citing 4,792,018 is not "the claims are dead" but "the patent is dead," because its 20-year term ran out on 2005-12-20 per the structured data, nearly seven years before the first IPR was even filed and more than twenty years ago as of today.

No proceedings to report — and I will not invent proceeding numbers

The instruction is explicit and I am honoring it: I found no PTAB proceeding on this patent and am therefore not populating the per-proceeding template. Fabricating an IPR number, a judge panel, or a claim-level disposition here would be worse than reporting an empty set.

What I searched and what I did not find:

• The canonical ODP "PTAB proceedings on file" block (authoritative) — empty.
• Web searches for the patent number paired with IPR/PGR/CBM terminology, and for the patent owner names (Checkrobot Inc., Uniquest, NewCheck, Productivity Solutions) in a PTAB context — returned only unrelated patents whose numbers happen to end in "018" (e.g., 6,744,018; 7,594,018; 8,911,479; 9,546,479) and the patent's own family/specification documents. Not a single hit tying 4,792,018 to a Board docket.
• No AIA trial petition, institution decision, FWD, or settlement/termination document was located.

Confidence and its limits: I could not load the PTAB E2E / Patent Trial and Appeal Board search UI directly and run a party-in-interest query; my absence finding rests on the ODP structured block plus search coverage. I would characterize this as high confidence, not cryptographic certainty. If you want the belt-and-suspenders check, run the patent number against PTAB E2E and CourtListener's docket search yourself before relying on it in a filing. Nothing I found contradicts the "no PTAB activity" default.

Why the absence is structural, not just luck. Two independent reasons explain why no AIA trial ever attached to this patent:

1. PGR is legally unavailable. Post-grant review is limited to patents whose claims have an effective filing date on or after 2013-03-16 (AIA § 6(c)(2)). This patent claims priority to 1984-07-09, with a 1985-06-12 filing date. It can never be a PGR candidate.
2. The assertion window closed before the IPR era opened. IPRs became available 2012-09-16; CBM review ran 2012-09-16 to 2020-09-16. Both require a real-world trigger — for CBM, the petitioner must have been sued for or charged with infringement. This patent's term expired 2005-12-20, so the demand letters that normally generate IPR petitions stopped arriving about seven years before the Board could hear one. An expired patent can still be IPR'd within the six-year damages lookback, but nobody had a financial reason to bother.

Strategic summary

Claim status: all 17 claims are UNTESTED — none canceled, none sustained, all expired. There is no IPR narrowing to map and no surviving-claim list to report, because no claim of 4,792,018 has ever been adjudicated by the PTAB. Do not let anyone tell you a claim was "confirmed" by the Board either — silence is not validation. Independent claims 1, 9, 11, 12, and method claims 14–17 all stand exactly as they issued on 1988-12-20, and they are all unenforceable for post-term conduct because the patent term ended 2005-12-20 (structured data: "Expired - Lifetime," anticipated expiration 2005-12-20). Practically: the patent is prior art and a historical citation, not an assertion asset.

Estoppel landscape: § 315(e)(2) estoppel is a non-issue, and that cuts both ways. Because no petitioner ever instituted an IPR, there is no § 315(e)(2) estoppel binding anyone on this patent — no prior-art ground was raised or reasonably could have been raised, so no party is estopped as to 4,792,018. That is cold comfort for a defendant, because a defendant's real problem is the 2005 expiration, not an estoppel shield they'd want anyway. Conversely, if you are evaluating this patent as part of a freedom-to-operate or invalidity posture against the later self-checkout patents (the Checkrobot descendants and the NCR/Symbol/Schneider families that cite it — see the 134 forward citations listed on the Google Patents page), 4,792,018 is available as § 102(b) prior art in full, unencumbered by any intervening IPR record.

Pattern signals: none, because there was never a Board docket to form a pattern. No serial petitioner, no multi-IPR campaign, no defensive aggregator (no Unified Patents involvement — the Unified petition that surfaced in my searches was against US 8,911,479 / 9,546,479, an unrelated patent family), and no patent-owner appeal to the Federal Circuit to trace. The ownership chain in the structured data tells the real story: Checkrobot → Omnicorp International (1990) → Uniquest (1994) → NewCheck (1994) → Silicon Valley Bank security interests (2001, terminated 2003–2004) → IBM (2005-04-25). IBM took the assignment in April 2005, roughly eight months before the term expired — a monetization-era transfer that never ripened into litigation, and certainly never into a PTAB campaign.

One thing I did not verify and you should. Ex parte reexamination is a separate track from AIA trials and is not reflected in the ODP "PTAB proceedings" block. Because this patent's specification and claims are heavily relied on as prior art by later self-checkout patents, it is worth confirming directly at the USPTO whether any ex parte reexam was ever filed against it — a certificate of reexamination would be reported on the patent's face and in the file history rather than in the PTAB data. I found no evidence of one, but I am flagging it as unchecked rather than asserting a negative I did not confirm.

Recommended next steps

• If you are a defendant facing assertion of US 4,792,018: the winning argument is expiration, not invalidity. Demand the plaintiff identify the accused conduct and its date. The structured data records the legal status as "Expired - Lifetime" with an anticipated expiration of 2005-12-20; any infringement theory must rest entirely on pre-expiration conduct within the § 286 six-year damages lookback, which now reaches back only to 2020-10-01 — well after the patent died. A claim chart for post-2005 conduct is facially defective.
• If you are assessing this patent as prior art against a later self-checkout patent (e.g., Mergenthaler US 4,779,706, Johnson US 4,787,467, or the Schneider self-checkout family that explicitly discusses it), quote the disclosure directly from the Google Patents full text — particularly the light-curtain shape measurement (FIGS. 4–5), the in-tunnel second UPC reader 130 (FIG. 6), and the EAS-tag verification routine (FIG. 9, steps 204–215). Those are the disclosures later patents cite.
• If you need to be certain about the empty proceeding set: verify at PTAB E2E and CourtListener before filing anything that turns on the absence. I state the absence with high confidence based on the ODP structured block and my searches, but I could not execute a party-in-interest docket query against the E2E interface directly, and I would rather tell you that than overstate it.
• If you were expecting an active proceeding: there is none to monitor. No institution deadline, no oral hearing date, and no statutory FWD due date exists for this patent, and I will not manufacture trial-stage milestones for a docket that was never opened.

I'll research the assignment record. Let me pull what I can from USPTO-linked sources.

US 4792018 — Assignment / Ownership Reconstruction

Patent: US 4,792,018 B1 — "System for security processing of retailed articles"
App: 06/742,757 · filed 1985‑06‑12 · granted 1988‑12‑20 · priority 1984‑07‑09 (CIP of Ser. No. 628,913, now US 4,676,343)
Sources actually available in this session: the authoritative Google Patents record at https://patents.google.com/patent/US4792018/en (including its Legal Events / Assignments module and its "Current Assignee" field). I was unable to retrieve live USPTO Assignment Center hit records for this patent — see the data-gap note under Assignment timeline.

Inventors

Unusual-pattern check — no mass-departure pattern. Contrary to the "all inventors exit within 12 months" precursor this task asks about, the record shows the opposite for at least one inventor: Steven J. Tilidetzke reappears as a named inventor on a later Checkrobot Inc. patent filed 1989‑03‑07 / granted 1990‑07‑10 (surfaced via Justia's CheckRobot Inc. assignee listing, https://patents.justia.com/assignee/checkrobot-inc). That is ~4 years after the US 4,792,018 filing, i.e. continued tenure rather than a 12‑month exit. David Humble also continues to appear as an inventor in the same technical space in the mid‑1990s (e.g. WO 1995/004491 A1 "Integrated automated retail checkout terminal," listed in this patent's own Cited‑By set), which is inconsistent with an inventor walk-out at filing.

Caveat: inventor employment is inferred from the face-of-patent assignee (no separate employment agreements were retrieved). I have not verified any individual inventor's departure date.

Original assignee

Checkrobot Inc. (original assignee named at grant; later records render it as "CHECKROBOT INC., A CORP OF DE" — a Delaware corporation, operating from Deerfield Beach, Florida).

• Primary line of business: automated/operator‑unattended supermarket self‑checkout — the "ACM" (Automated Checkout Machine) system. This is the archetypal self‑checkout counter patent family; the parent US 4,676,343 is the foundational Checkrobot document.
• Did it ship a product embodying the claims? — Yes. This was not a paper-only portfolio. Contemporaneous trade coverage (Computerworld, Internet Archive scan, https://ia802304.us.archive.org/18/items/computerworld2435unse/computerworld2435unse.pdf) records that Pick 'n Save stores in suburban Milwaukee installed a self‑checkout system developed by Checkrobot, Inc. of Deerfield, Fla., and that Kroger piloted then dropped the Checkrobot system in two suburban Atlanta stores ("too impersonal"). A commercially deployed, then partially de‑installed, product is a classic operating-company profile.
• Current status: the recorded chain shows Checkrobot divesting this patent in 1994 (see timeline), after a 1990 collateral security interest. I could not independently verify whether the disposition was a bankruptcy/Chapter 11 asset sale, an orderly wind‑down, or a business-line sale — and I will not assert bankruptcy without a filing document. What the record does show is that the self‑checkout business continued under successor names (Uniquest → NewCheck / Productivity Solutions → IBM), which is more consistent with a business transfer than a portfolio fire-sale.

Assignment timeline

Data-gap disclosure (important): Google Patents' Legal Events module does not expose reel/frame numbers, execution dates, or the correspondent/recording attorney. I was not able to complete a live USPTO Assignment Center lookup for patent 4792018 / app 06/742,757 in this session (assignmentcenter.uspto.gov does not expose its contents to the search tools available here). Therefore I cannot supply reel/frame identifiers or correspondents for this chain, and I will not invent them. The dates below are the dates surfaced by Google Patents' reassignment events; Google does not label them unambiguously as execution vs. recordation dates, so treat each as "date surfaced," not as a verified execution date. To complete the reel/frame and correspondent columns, query https://assignmentcenter.uspto.gov/ (or the mirrored https://assignment.uspto.gov/patent/index.html) by patent number 4792018 or application 06/742,757.

1. 1990‑09‑21 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Security Interest ("SECURITY INTEREST (SEE DOCUMENT FOR DETAILS)")
   • Assignor: Checkrobot Inc., a corp. of Delaware
   • Assignee: Omnicorp International B.V.
   • Correspondent: not retrievable from the source used
   • Context: Securitization — collateral grant over the patent to a Dutch (B.V.) financier; this is a lien, not an ownership transfer.

2. 1994‑08‑19 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Assignment of Assignors' Interest
   • Assignor: Checkrobot, Inc.
   • Assignee: Uniquest, Incorporated
   • Correspondent: not retrievable from the source used
   • Context: Asset/business transfer out of the original operating company — the first true ownership change.

3. 1994‑11‑15 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Assignment of Assignors' Interest
   • Assignor: Uniquest, Incorporated
   • Assignee: NewCheck Corporation (address on record: 10475 Fortune Parkway)
   • Correspondent: not retrievable from the source used
   • Context: Transfer-to-successor / internal reorg of the self-checkout business — Uniquest held the asset for under three months before passing it to NewCheck.

4. 2001‑08‑20 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Security Interest
   • Assignor: NewCheck Corporation d/b/a Productivity Solutions, Inc.
   • Assignee: Silicon Valley Bank
   • Correspondent: not retrievable from the source used
   • Context: Securitization — venture-debt collateral over the portfolio.

5. 2003‑10‑23 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Security Interest Termination
   • Assignor: Silicon Valley Bank
   • Assignee: NewCheck Corporation d/b/a Productivity Solutions, Inc.
   • Correspondent: not retrievable from the source used
   • Context: Lien release / loan payoff.

6. 2004‑04‑19 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Release
   • Assignor: Silicon Valley Bank
   • Assignee: New Check Corporation
   • Correspondent: not retrievable from the source used
   • Context: Second lien-clearing filing — note the assignee name is rendered with a space ("New Check") versus "NewCheck" elsewhere, suggesting a possible corporate name change between the two SVB filings. Unresolved.

7. 2005‑04‑25 (surfaced) / recorded date not exposed — Reel/frame not retrieved

   • Conveyance: Assignment of Assignors' Interest
   • Assignor: New Check Corporation d/b/a Productivity Solutions, Inc.
   • Assignee: International Business Machines Corporation
   • Correspondent: not retrievable from the source used
   • Context: Business-line acquisition — consistent with IBM's 2005 acquisition of the Productivity Solutions/NewCheck self-checkout business; the patent went with the business.

Post‑chain event (not an assignment): 2005‑12‑20 — Anticipated expiration. The 17‑year pre‑URAA term from the 1988‑12‑20 grant ran out; the patent is now Expired – Lifetime. Note that IBM's 2005 acquisition of the asset was followed by expiry of the patent eight months later, a very narrow assertion window if one had ever been contemplated.

Google Patents also lists a "Current Assignee" set of "Uniquest Inc / International Business Machines Corp / NewCheck Corp." Google itself disclaims accuracy of that field ("listed assignees may be inaccurate"), and it appears to be a roll-up of historical assignees rather than current title. The last recorded ownership event is the 2005‑04‑25 assignment to IBM.

Timeline diagram

timeline
    title Ownership of US 4792018
    1984 : Priority date 9 July 1984
    1985 : Application filed 12 June 1985
    1988 : Patent issues 20 December 1988
    1990 : Security interest to Omnicorp International
    1994 : Assigned to Uniquest Incorporated
         : Assigned to NewCheck Corporation
    2001 : Security interest to Silicon Valley Bank
    2003 : SVB security interest terminated
    2004 : SVB release recorded
    2005 : Assigned to IBM
         : Patent term expires 20 December 2005

NPE / troll-pattern signals

Verdict

Defensive / non-asserting.

The recorded chain runs Checkrobot Inc. → Uniquest Inc. (1994‑08‑19) → NewCheck Corp. (1994‑11‑15) → Silicon Valley Bank liens (2001‑08‑20, terminated 2003‑10‑23 and released 2004‑04‑19) → International Business Machines Corp. (2005‑04‑25), and no entity on the required NPE list appears anywhere in those seven events; the original assignee was a genuine operating company that shipped and installed the claimed self-checkout system (Pick 'n Save, Kroger). The patent then expired 2005‑12‑20, eight months after the last transfer, so it is now unassertable irrespective of title. Caveat on the label: the chain does not terminate at one of the named defensive aggregators, so "Defensive / non-asserting" here reflects an operating-company endpoint plus term expiry, not an RPX/AST/LOT acquisition — and signal #3 (repeat correspondent) remains unverified, which is the one check that could still move the needle.

Verify at: https://assignmentcenter.uspto.gov/ (search patent 4792018 or application 06/742,757) and the mirror https://assignment.uspto.gov/patent/index.html; patent face record at https://patents.google.com/patent/US4792018/en.

I'll search for the authoritative citation/reference list for US 4,792,018, since the patent text supplied was truncated mid-way through the cited-references table.

Let me pull the citation list from other databases (FreePatentsOnline, Espacenet) and verify the remaining cited references.

Let me try to retrieve the complete cited-references table and verify individual references.

Prior-Art Analysis — U.S. Patent No. 4,792,018

Task scope: Identify the prior art cited on the face of U.S. 4,792,018 and assess § 102 anticipation potential claim-by-claim.

Important limitation stated up front (no fabrication): The authoritative full text supplied for this analysis contains the front-page citation table under the heading "Citations (13)", but the text I was given is truncated mid-way through the eighth entry (it cuts off at "US3832530A … Westinghouse Electric Corp — Object"). I was able to confirm entry 8's full title by search (Object identifying apparatus), but I could not retrieve the remaining 5 of the 13 cited references before my tool budget was exhausted. I therefore analyze the 8 identified citations plus the prior art discussed in the specification's own Background section, and I explicitly do not guess at the missing five. This should be read as a partial, high-confidence-for-what-it-covers analysis, not a complete § 102 clearance.

1. Sources searched (and false positives excluded)

Excluded false positives (per the strict-literal rule, not conflated with this patent): JP 4792018 B2 (a Japanese grant cited in EP1982607B1), the string "4792018" appearing in a Brazilian corporate registry, and US 7,920,018 (Advanced Memory Technologies). None is US 4,792,018.

Cross-check note: A search hit for US 5,239,167 (Kipp, "Checkout system") shows both US 2,919,851 (Otis) and US 4,792,018 in its "Referenced Cited" list — useful confirmation that the Otis patent and the '018 patent sit in the same recognized art cluster.

2. The identified cited references (entries 1–8 of 13)

Dates below are exactly as recorded in the "Citations" table of the supplied full text (Google shows a priority/filing date and a publication date column).

Anomaly flag (reported literally, not auto-corrected): Google's title field for FR 899,018 A reads "Children's romper." That is plainly irreconcilable with the citation context of an electro-mechanical checkout-security patent, and I judge it to be a metadata error in the source — but I cannot state what FR 899,018 actually discloses, so I map it to no claim.

3. Per-reference analysis and § 102 mapping

Anticipation standard applied: a single reference anticipates only if it discloses every limitation of the claim, arranged as in the claim. Where a reference teaches only one element (e.g., a code reader, or a conveyor), it is cited for that element and does not anticipate the claim as a whole. All full-text readings are provisional where noted — my searches returned front-page/index data, not complete specifications, for entries 1–4 and 7.

(1) FR 899,018 A — 1945-05-15

• Description: Unverifiable. Google's title field returns an unrelated garment title; no reliable disclosure available to me.
• § 102 potential: None that can be responsibly attributed. Cannot be mapped to any of claims 1–17 without reading the actual document. Priority verification item.

(2) US 3,436,968 A — Fairbanks Morse, "Processing control system" (filed 1965-02-11; issued 1969-04-08)

• Description: Fairbanks Morse was a scale/weighing and materials-handling house; the title indicates a control system for a processing (conveyed-article/weighing) line. I did not verify the body text.
• § 102 potential: Element-level only — bears on the conveyor means of claims 1(b), 9(b), 11(b), 12(b) and, at most, on the weight-measurement aspects of claims 3 and 17. No code-correlated accept/reject logic, so it does not anticipate any claim.

(3) US 3,511,372 A — Kantarian, "Article handling means, systems, and devices" (filed 1968-02-26; issued 1970-05-12)

• Description: Article handling/transport art (conveying and routing of articles). Body text not verified.
• § 102 potential: Element-level only — conveyor/transport means. No anticipation of any claim.

(4) DE 2,114,676 A1 — Bendix, "Optical label reader" (filed 1970-04-09; published 1971-10-21)

• Description: Machine-reading of labels/codes by optical means — general code-reader art.
• § 102 potential: Element-level only — the "code reader means" of claims 1(a), 9(a), 11(a), 12(a). No anticipation of any claim (no conveyor security logic, no stored-value comparison, no EAS, no sentry).

(5) US 3,681,570 A — Zellweger Uster AG, "Method and apparatus for identifying articles" (filed 1969-03-26; issued 1972-08-01) — the strongest cited reference

• Description (as characterized by the '018 specification itself, which names it "Abt U.S. Pat. No. 3,681,570"): a system that accepts or rejects on the basis of comparing a measured article characteristic against a preassigned value therefor, correlated with article identification code; articles examined per article; accepted articles are conveyor-transported to a secured container inaccessible to the customer until payment (no human bagging); rejected articles are diverted to a second (rejection) conveyor and returned to the customer.
• § 102 potential:
  • Claim 1 — close, but no. The '570 discloses the "code-correlated stored value → measure → compare → accept/reject" core of claims 1(d)(I)–(III). It does not disclose claim 1(c)'s sentry "jointly indicative of entry … and of a first measurable characteristic" (the single-sensor inlet function), nor the reversing entry conveyor of the '018. Element-by-element, claim 1 is not met.
  • Claim 3 — no (no light-curtain shape sensing + weight combination as claimed).
  • Claims 14(I)/15(I) — the "failure of correspondence of measured characteristic with a stored predetermined value → reject and return" step is squarely taught by '570's rejection conveyor. Claims 14 and 15 as a whole, however, additionally require an EAS practice (14(II)(B), 15(II)) that '570 lacks, so neither method claim is anticipated.
  • Role in the record: primary § 103 base, not a § 102 reference.

(6) US 3,681,571 A — Zellweger Uster AG, "System for identifying articles" (filed 1969-04-21; issued 1972-08-01) — co-strongest

• Description: Companion to '570 (named "Strohschneider" in the '018 specification), same characterization: per-article identification-code-correlated characteristic comparison with accept/reject.
• § 102 potential: Identical mapping to entry (5) — element-level relevance to claims 1(d), 3, and the "(I)" branches of claims 14/15; no anticipation of any claim as a whole. Primary § 103 base.

(7) US 3,734,286 A — L. Simjian, "Article handling apparatus" (filed 1971-12-27; issued 1973-05-22)

• Description: Article-handling apparatus (L. Simjian was prolific in automated merchandising/banking equipment). Body text not verified.
• § 102 potential: Element-level only — conveyor/article-handling and possible sensing means. No anticipation of any claim.

(8) US 3,832,530 A — Westinghouse Electric Corp, "Object identifying apparatus" (continuation filed 1972-01-04; issued 1974-08-27) — the closest cited art to the EAS claims

• Description (verified from the reference's own front matter and definitions on Google Patents): objects (mail sacks/baggage on a conveyor; the specification also contemplates "vehicles or persons passing into a military compound or other secure area") carry a label with a solid-state circuit — an antenna/coil plus counter means (interrupted-ring counters or shift registers) — that is powered by and absorbs energy from a homogeneous electromagnetic field as the object moves through a region; the absorbed-energy pattern is decoded into an identifying code. Its cited-of-record references are 3,299,424 (Vinding, 1/1967) and 3,752,960 (Walton, 8/1973).
• § 102 potential:
  • Claims 4, 7, 9(f), 12(d), 14(II)(B), 15(II) — closest cited art, but no anticipation. '530 discloses electromagnetic (non-optical) detection of a device-borne tag on a conveyed object, on a conveyor, including in a secure-area entry context — i.e., the physical genus of an "EAS detection means … determining whether or not an article … is EAS-tagged." But '530 decodes an identity code; it does not disclose detecting the presence/absence of a tag or comparing that against a per-UPC stored indication of whether the article should be tagged. The '018's distinctive EAS logic (store, per article ID, whether it should be tagged; compare to detection) is absent.
  • Claim 12 — requires the controller to store an EAS-should-be-tagged indication correlated with the article identification code and compare it to the detector output. '530 has no such stored-indication/comparison step. No anticipation.
  • Role in the record: best § 103 starting point for the EAS-elements of claims 4/7/9/12/14/15.

4. Prior art discussed in the specification but not shown in the identified citation rows

US 2,919,851 — Otis, issued January 5, 1960 (expressly cited in the '018 Background)

• Description (from the '018 text): a machine-discernible code is assigned to each article, as is a machine-discernible weight indication within a tolerance range; code and weight indications are discerned for selected articles and weights totalized with tolerance; the customer places the shopping bag in a restricted area on a scale; if measured bag weight corresponds with the totalized weight, the transaction is not rejected; if it does not correspond, the customer is directed to the store manager (no automatic conveyor return). Also provides machine-discernible price indication and a printed record with price totalization.
• § 102 potential: Relevant to claims 3 and 17 (weight measurement compared against a stored value correlated with article identity) and to the general "predetermined value correlated with identification code" concept in claims 1(d)(I), 9, 11, 12. No anticipation — Otis is transaction/bag-level, uses human-manager intervention rather than conveyor reversal, and discloses no sentry, no in-zone second code read, no EAS, and no self-compiling target store.
• Note on the citation list: the '018 cites Otis in the specification body, and Google's citation table (as far as I can read it) runs in ascending date order from FR 899,018 A (1945) straight to US 3,436,968 A (1969) — so Otis does not appear among entries 1–8. Whether it appears among the five missing entries (which would have to break the apparent date ordering) is unverified.

US 3,681,570 / US 3,681,571 (Abt / Strohschneider, Zellweger)

These are cited both in the specification's Background and as citations (5) and (6) above — no duplication issue; treated consistently.

5. § 102 summary matrix (claims vs. identified citations)

Bottom line on anticipation: On the record available to me, none of the eight identified citations anticipates any of claims 1–17 as a whole. Every one of them is element-level art: code reading (Bendix '676), article handling/conveying (Kantarian '372, Fairbanks Morse '968, Simjian '286), for-identification label detection (Westinghouse '530), and per-article code-correlated accept/reject comparison (Zellweger '570/'571). The '018's novelty rests on combinations no single citation supplies — the dual-function entry sentry, the reversing entry conveyor, the in-zone second code read, the EAS should-be-tagged-vs-detected logic, and the self-compiling SETUP store (H1/H2/H3 + weight) of Fig. 13. The realistic threat to these claims is a § 103 combination, principally: Zellweger '570/'571 (comparison core) + Otis '851 (code-correlated weight) + the '913 parent's light-curtain sentry + a 3,832,530-type electromagnetic tag detector.

6. Documents in the record that are not prior art (flagged to prevent miscitation)

• US 4,676,343 ("Self-service distribution system," Ser. No. 628,913, filed 1984-07-09, issued 1987-06-30) — the CIP parent of this patent and the source of the light-curtain/secured-zone disclosure. It is not § 102 art: it issued after the '018's 1985-06-12 filing date and shares the priority chain. It is the § 112 written-description support, not a reference.
• "Cited By (134)" and "Families Citing this family (67)" entries (e.g., US 4,692,747 Article security system, 1987; DE 3,913,162 A1, 1989; US 4,909,356; US 5,239,167; US 5,594,228; US 5,340,970; US 5,426,282) all post-date the '018 and are later art — they show the '018 was influential, not that it was anticipated. Do not treat them as references against it.
• US 5,239,167 (Kipp) is useful only as evidence of the art cluster (it cites both Otis '851 and the '018).

7. Confidence, gaps, and what would close them

1. High confidence in the eight identified citations' bibliographic data (taken from the authoritative supplied full text and confirmed by independent search for entries 5, 6 and 8).
2. Unverified (explicit gap): the 5 remaining of the 13 front-page citations were truncated in the supplied text and I could not retrieve them before exhausting my tool budget. My "no anticipation by any cited reference" conclusion is therefore provisional as to those five. If one of them is, e.g., a 1970s electronic-article-surveillance or automatic-weighing-checkout patent, the EAS-claim analysis could change.
3. Unverified full texts: I retrieved front-page/index data and referenced descriptions (not complete specifications) for US 3,436,968; US 3,511,372; DE 2,114,676; FR 899,018; and US 3,734,286. Element-level assignments for those five are title-and-context-based and should be confirmed against the documents.
4. Recommended next steps: (a) pull the complete "References Cited" list from USPTO Patent Public Search / PatentCenter for US 4,792,018 to obtain entries 9–13; (b) read the full text of US 3,681,570 and US 3,681,571 — they are the load-bearing references; (c) resolve the FR 899,018 title anomaly; (d) confirm whether Otis US 2,919,851 is in fact one of the 13 front-page citations or only a specification-discussed reference.
5. No contradiction was found between this analysis and the previously generated bibliographic/claims/litigation sections; item 3 of that section's uncertainty notes (the truncated citation list) is confirmed and quantified here as 8 of 13 identified.

I'll ground this obviousness analysis in the prior art actually cited on the face of US 4,792,018, and verify the truncated citation entries before relying on them.

US 3,832,530 is confirmed as Westinghouse's "Object identifying apparatus" (RF-coded label / energy-absorption identification). Let me verify the remaining truncated citations.

US 3,436,968 is confirmed as a conveyor-based height/width/length + weight processing system. Let me characterize the two non-English citations and the Otis reference.

Obviousness Analysis — U.S. Patent No. 4,792,018 under 35 U.S.C. § 103

Analyst note on scope: Per the prior-art directive, this analysis uses (i) the background references identified in the '018 specification itself — Otis US 2,919,851 and the Zellweger '570/'571 patents (US 3,681,570 / US 3,681,571) — and (ii) the 13 references in the "Citations" section of the Google Patents record (https://patents.google.com/patent/US4792018/en). I verified three of the previously-truncated citation entries by direct retrieval (US 3,832,530, US 3,734,286, US 3,436,968). See §7 Caveats for what remains unverified.

I do not repeat the bibliographic and claim-summary sections produced earlier; claim language is cross-referenced to those sections.

1. Governing legal framework and the threshold issue: which priority date applies

1.1 PHOSITA

A person having ordinary skill in the art as of 1984–85 would be a systems engineer in automated point-of-sale / retail checkout equipment, with working familiarity with: (a) optical (UPC/bar-code) and alphanumeric scanning; (b) conveyor drive and diversion controls; (c) load-cell and platform weighing; (d) photoelectric/light-curtain sensors and multiplexed LED-photocell arrays; and (e) electronic article surveillance (magnetic, RF, and electromagnetic energy-absorption marker systems). This is a narrow, mature, and highly integrated art: the individual subsystems were all commercially available before 1960.

1.2 The CIP date-splitting problem — the single most important § 103 vulnerability

The '018 is a continuation-in-part of application 06/628,913, now US 4,676,343 (filed 1984-07-09). Standing alone:

• Subject matter supported by the '913 parent (code reader + conveyor + secured zone + entry/exit light curtains + weight comparison + conveyor reversal + interactive display) gets the 1984-07-09 date.
• Subject matter first added in the '018 gets only the 1985-06-12 filing date. On the face of the specification, the following appear to be '018-only subject matter:
  • Derivation of article shape/height from the entry light curtain (Fig. 5's counter 152 + LED/PC multiplexers 148/158, "initial vertical LED-PC pair in communication," indicating height; shape sensor 108 and shape comparator 114).
  • The EAS tag-presence/absence logic (EAS detector 134, steps 206–215 of Fig. 9).
  • The SETUP/self-compiling data base (Fig. 13, steps 258–272, including H1/H2/H3 height storage).
  • The in-zone second UPC reader 130 with X-Y drive 172 (Fig. 6) is arguably new matter as to its implementation, though the generic concept of code verification against a stored value is squarely in the parent.

Consequence: for these claim groups the critical § 102(b) cutoff is 1984-06-12, and § 102(a)/(e) art published/filed between 1984-07-09 and 1985-06-12 also becomes available. Every date-bearing reference cited on the '018 face (1943–1974) predates both cutoffs by a decade or more, so the priority dispute does not rescue these claims. This should be pleaded expressly in any obviousness contention rather than left implicit.

1.3 The parent is not prior art

US 4,676,343 must not be used as a § 102/§ 103 reference against the '018: common inventors, common assignee (Checkrobot Inc.), and the same inventive entity. It is useful only as evidence of the state of the art and of the ordinary artisan's knowledge at the shared priority date — precisely the role the '018 specification itself assigns to the '913 application ("to which incorporating reference is hereby made").

1.4 Methodological point on the record

The Google Patents page lists 134 "Cited By" documents. Those post-date the '018 and are not prior art — they are later art (e.g., US 5,595,228; the NCR self-checkout series; DE3913162A1 to Checkrobot). They may be cited only as objective evidence of what the art later regarded as routine (e.g., that reconciling a scanned code against a detected physical/tag attribute is standard), not as § 103 references. Only the 13 "Citations" plus the background references are proper.

2. Verified characterizations of the principal references

3. Ground-by-ground obviousness analysis

Ground 1 — Claims 1, 2, 3: Otis '851 in view of Zellweger '570/'571, further in view of Simjian '286 and Fairbanks Morse '968

Claim 1 element chart

Why the combination is obvious (motivation, articulated as of 1984–85):

1. Same problem, same art, same solution class. Otis and Zellweger are both unattended-checkout fraud-deterrence systems whose entire disclosed security logic is "sense an article attribute → compare to a value pre-assigned to the scanned code → route the article." Claim 1 is that logic. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) — where the references address the same problem with the same mechanism, combination is not inventive.
2. The dimensional sensor is a known equivalent of the weight sensor. Otis's machine-discernible attribute is weight; Fairbanks Morse '968 teaches that height/width/length of an article moving on a conveyor can be measured and combined with weight by data processing. Substituting or adding a dimensional attribute for a weight attribute is the substitution of one known article-attribute sensor for another, to obtain the predictable result (a second independent fraud check). KSR; In re Fout.
3. Dual use of the existing entry sensor — the strongest motivation for the "jointly indicative" limitation. Simjian '286 already discloses that the entry/position sensor beam is used as (i) a gating/entry signal and (ii) a trigger for downstream action. Once a stepping light curtain is used at the inlet (as the '913 parent does for security), the artisan has a cost and packaging incentive to take the same sensor's outputs and interpret the highest interrupted beam as height — zero added hardware, one sensor, two data products. This is precisely the kind of "use of a known technique to improve a similar device in the same way" that KSR holds obvious.
4. The reference-count incentive. Both Otis and the '018 specification itself recognize that each added independent attribute increases the number of independent bases for rejection. The '018's own fraud example (scan the cheap wine bottle's UPC, deposit the expensive bottle) demonstrates that weight alone is defeatable when the bottles are similarly sized and weighted — which is a motivation to add shape, supplied by the admitted problem, not by the inventor.

Claim 2 (self-compiling store). Rendered obvious by claim 1 in view of Fairbanks Morse '968's data-processing record generation, and by the ordinary artisan's knowledge that a retailer's UPC/price file does not contain weight or shape data. The artisan must therefore either (i) manually enter attributes (labor cost, error), or (ii) have the system observe and store them. Choice (ii) is the routine engineering alternative. The H1/H2/H3 multiple-height storage in the '018's Fig. 13 is not inventive: the specification itself concedes that a can presents length or diameter, and a box presents length or width or height — a finite, predetermined, and fully predictable set of orientations. Storing the closed set of expected presentations is a mechanical accommodation, not invention.

Claim 3 (shape, plus weight sensing and comparison). Obvious for the reasons above: Otis supplies machine-discernible weight with tolerance comparison; Fairbanks Morse '968 supplies dimensional measurement in combination with weight data processing; Zellweger supplies the accept/reject routing. Adding a scale under the entry conveyor (a load cell under a conveyor is a routine expedient, and the '913 parent places a weight sensor under conveyor 12) yields the claim-3 system with no new principle of operation.

Expected result / no unexpected results. Each check operates independently on the same article at the same time; the combination produces the predictable additive benefit (more fraud modes caught) with no observed interaction. No teaching away: neither Otis, Zellweger, Simjian, nor Fairbanks Morse disparages dimensional or weight verification.

Ground 2 — Claim 4 (EAS determination + per-code stored tag expectation): Ground 1 in further view of Westinghouse '530

Claim 4 adds: "EAS detection means for determining whether or not an article in said security zone is EAS-tagged," with the control means storing, per article identification code, whether the article should be tagged, and operating the conveyor on stored-indication vs. detection.

Where the elements come from:

• "EAS detection means… determining whether an article in the zone is tagged": Westinghouse '530 discloses an article-mounted circuit that absorbs electromagnetic field energy, producing "a succession of pulses… which serve as a code to identify the object," usable for objects passing into "a military compound or other secure area." That is a detector positioned to interrogate an object moving through a region, producing an electrical output indicative of the presence and identity of an object-borne electronic circuit. The detector half of claim 4 requires no more.
• Storing, per article ID code, whether the article should be tagged: this is Otis's own architecture — a machine-discernible attribute assigned to and stored against the article's code. The only difference is that the stored attribute is "is tag expected?" (binary) rather than "what weight?" The claim thus recites the substitution of a binary tag-state field for the numeric weight field in an otherwise identical lookup-and-compare control loop.
• Selective tagging is admitted prior practice. The '018 specification itself frames EAS as existing store infrastructure ("the pedestrian passage about the system may include an EAS… to defeat circumvention"), and describes selective tagging of expensive articles as "one approach" — i.e., as known. Using the system to verify a segregation rule that the store has already adopted by economic necessity (tags cost money, so only some articles get them) is a predictable application of a stored expectation field.

Motivation: the store already deploys EAS at the doorway; ignoring the tag state inside the secured lane wastes a sensor and leaves the wine-bottle-substitution fraud open. The artisan, seeking "plural bases for rejection," would naturally read tag presence in the lane and compare it against the store's own tagging policy. This is a combination of prior-art elements according to known methods to yield no more than predictable results — obvious as a matter of law under KSR.

Ground 3 — Claim 6 (second code reader inside the zone): Ground 1 in further view of Simjian '286 and Westinghouse '530

Claim 6 adds "additional code reader means in said security zone" and comparison of the inside read to the outside read.

Motivation is the strongest of any in this patent. Zellweger's entire premise is that accepted articles travel to a secured container and rejected ones to a rejection conveyor; the security architecture of the field is physically built around an in-zone region. Simjian '286 discloses the code-reading function performed on the article while it is inside the enclosed sensing chamber 30 — i.e., reading a code at an in-zone station, in a conveyorized flow. Adding a second read inside the zone to verify the same code (a redundant read, or a re-read after physical transfer) is the classic redundancy/verification step: the field's known answer to "did the physical article match the scanned code?" is to sense something about the physical article after scanning (Otis: weight; Zellweger: preassigned characteristic; the '913 parent: weight). A second code read is a third known alternative attribute sensor (§ 103 prima facie on the "known alternatives" rationale). Reading a code at more than one station along a conveyor path is itself routine in material handling, and Bendix's optical label reader reflects that a label can be read at a designated station.

Note on drafting scope: claim 6 is a dependent-form claim that adds the second reader without requiring the characteristic comparison, so it reads on any two-station code-verification checkout lane. That breadth increases its exposure to a § 103 rejection over Otis + Simjian + Zellweger.

Ground 4 — Claim 9 (dual read and EAS): Grounds 2 + 3 combined

Claim 9 is the union of the dual-reader basis and the EAS basis in one system, and it is the claim group most exposed because it omits any measured-physical-characteristic comparison — the very feature that the '018 specification touts as the improvement over the '913 parent. With that feature removed, claim 9 reduces to: known unattended-checkout architecture (Otis/Zellweger) + known in-zone code station (Simjian) + known article-borne coded-circuit detector (Westinghouse '530) + a per-code expectation table (Otis). The combination of two independent fraud checks known in the same art ("plural bases for the rejection of articles") is the express object of the invention and an obvious design goal; aggregating known checks does not create patentability.

Claim 10 (self-compiling store added to claim 9) fails with Ground 1's claim-2 reasoning. Note claim 10's reference to "sentry means" although claim 10 depends from claim 9, which recites no sentry means — a § 112 antecedent-basis defect that also narrows the practical force of the "self-compiling store" recitation.

Ground 5 — Claims 11 and 13 (self-compiling system, independent): Otis '851 / Zellweger '570/'571 in view of Fairbanks Morse '968

Claim 11 removes the sentry and the entry-detection limitation entirely. It recites: reader + conveyor + secured zone + sensing means for a measurable characteristic + control means that compiles its own store of characteristic-vs-code values by processing reader and sensor outputs, then compares and routes.

• The compile-then-compare loop is '968 plus Otis. Fairbanks Morse '968 discloses measuring an article's dimensions on a conveyor and processing those data in a computer to produce output control information, optionally in combination with weight; Otis discloses storing per-code expected attributes and comparing. Associating a measured attribute with the code just read from the same article — i.e., teaching-by-example — is the most natural way to populate an attribute file, and requires only ordinary record-handling that '968's data-processing means already contemplates.
• Motivation, explicitly available: the '018 specification concedes the alternative ("the in-place facility data base may be used in the absence of SETUP") and asserts that SETUP gains "an independence from the UPC-related data base of the facility." Independence from an external file, and elimination of labor to populate it, is a classic design/cost motivation of the kind KSR treats as sufficient.
• Claim 13 adds the self-compiling store to an EAS-based system (Ground 2 + Ground 5). Same reasoning; the additional presence of an EAS detector does not change the analysis.

Ground 6 — Claim 12 (EAS-only independent system): Otis '851 in view of Westinghouse '530 (in further view of the '913 parent's admitted EAS doorway)

Claim 12 is, on its face, the broadest claim in the patent: reader + conveyor + secured zone + EAS detector producing a signal + a control means that stores, correlated with the article ID code, whether the article should be tagged, compares the detector signal to that stored indication, and routes the conveyor accordingly. It contains no weight, shape, dimensional, or second-scanner limitation.

Obviousness follows from:

• Westinghouse '530 (detector interrogating an object-borne circuit in a secure-area passageway, producing an identifying output) supplies (d);
• Otis '851 supplies (a), (b), and the entire "(e) store a per-code attribute → compare sensed value → accept/reject" control architecture, with the stored attribute being a binary tag-policy field instead of a weight value;
• Zellweger supplies the secured-zone conveyor routing and reject path;
• the '018 specification's own admission supplies the motivation (the EAS doorway exists to "defeat circumvention"; selective tagging of expensive articles is "one approach").

A control system that validates a detected state against a stored policy for the identified item is the core of Otis's teaching, and applying it to a tag state is not a new principle of operation. Expect this claim to be the lead § 103 target.

Ground 7 — Claims 14, 15, 16, 17 (the method claims): Grounds 1–6 applied as method steps

The method claims recite the operating method of the systems above and stand or fall with them:

• Claim 15 (read code → place in secured zone → measure characteristic → EAS-examine → reject on characteristic mismatch or EAS failure) = Otis/Zellweger's compare-and-route method (Ground 1) + Westinghouse '530's tag interrogation (Ground 2), performed on the same article during the same in-zone transit.
• Claim 14 adds the second in-zone code read (Ground 3). Note the drafting quirk already flagged in the prior sections: claim 14's condition II is conjunctive ("A. failure of correspondence … and B. failure of such article to pass such EAS practice"), which on its face mandates rejection only when both the code mismatch and the EAS failure occur. That conjunctive branch is narrower than the system claims, but condition I of claim 14 and the disjunctive "or" between I and II preserve a broad rejection trigger (any single characteristic mismatch alone suffices) — so claim 14 remains substantially covered by Ground 1's method reasoning.
• Claim 16 (characteristic = article shape) is expressly met by Fairbanks Morse '968's conveyor-based height/width/length measurement of moving articles, in view of Otis/Zellweger's comparison-and-routing.
• Claim 17 (add weight measurement and weight-mismatch rejection) is met by Otis '851's machine-discernible weight-with-tolerance and by '968's express combination of dimensional data with weight — the claim recites the additive aggregation of two known checks, nothing more.

Method claims drawn to the ordinary operation of an obvious apparatus carry no independent patentable weight. In re King, In re Kuehl line of reasoning.

4. Consolidated grounds table

5. Anticipated patentee rebuttals and why they likely fail

1. "Non-analogous art / different field" — Unavailable. Simjian '286 ('article handling on a conveyor with optical code reading and an in-zone sensing chamber'), '968 (dimensional measurement of conveyed articles), and '530 (object identification in a secure-area passage) are all in material-handling/identification arts in which the '018 itself claims; the '018's field is the combination of exactly these functions. No reference is more than a step removed.
2. "Improper hindsight / no motivation to combine" — The motivation is documentary and pre-dating: the '913 parent frames generic weight comparison as insufficient for "satisfactory commercialization"; the '018 specification itself supplies the fraud scenario (wine bottles) that motivates a second and third attribute; Zellweger's secured-container architecture motivates in-zone sensing. Motivation here is not judicially hypothesized — it is recited in the applicant's own specification.
3. "The self-compiling SETUP mode is unexpected" — Weak. The result (an attribute file populated from observation) is precisely the expected result of observing articles and recording what is observed; the H1/H2/H3 scheme is dictated by the finite orientation set the specification itself enumerates.
4. "Teaching away" — No reference criticizes or discourages dimensional or tag-based verification. Otis's tolerance ranges, if anything, invite tighter supplementary checks.
5. Secondary considerations (to be tested, not conceded): commercial success of the Checkrobot/Productivity Solutions self-checkout installations, and the recorded chain to IBM (2005) and to UniQuest/NewCheck, would be the patentee's likely § 103 counterweight. Any such showing requires a nexus to the claimed invention; given that claim 12 and claim 9 read on generic code-reader + secured-lane + tag-detector combinations, a court is likely to find the commercial-success evidence tied to the product as a whole (and to the '913 parent's weight/light-curtain architecture) rather than to the '018's asserted improvements. This should be explored in discovery but is not a bar to the grounds above.

6. Where the record is thinnest for the challenger

• Claim 1's "jointly indicative" limitation is the most technically awkward element to meet cleanly. The best mapping is Simjian's single inlet beam (entry) combined with '968's dimensional measurement. A patentee will argue that a photoelectric beam (entry trigger) is not a height-measuring sensor, and that '968's dimensional arrays are used for volume/billing, not security. Counter with: (i) the '913 parent's own light curtain is admitted art for the zone inlet, and the artisan's inference of height from the interrupted beam index is a conversion, not a new component; (ii) analogous-use is irrelevant under KSR; and (iii) '968's express combination of dimensions with weight for a control output places it in the security-check use case by reasonable extension.
• Whether the second in-zone reader's X-Y aiming implementation (Fig. 6, alphanumeric recognition unit 166, drive 172, links 176/178) adds anything. Under § 103 it does not; aiming a scanner at a located label is routine and Bendix's optical label reader is evidence of label-reading at a station.

7. Caveats and uncertainties (stated candidly)

1. The "Citations" list is truncated in the supplied full text at its last visible entry, US 3,832,530 (Westinghouse). It is therefore possible that the complete examiner-cited set includes one or more additional references (the FR 899018 A entry at the top of that list suggests a longer, mixed-language set). The ground set above may be incomplete as a matter of record, though it is complete as to the references I can independently verify.
2. FR 899018 A should not be used. Its Google-Patents title ("Children's romper" / "Barboteuse pour enfants") is facially incongruent with this technical field and appears to be a record-mapping error. I have excluded it. Verification against the INPI/French original would be required before any reliance.
3. I retrieved and read only three of the cited references directly (US 3,832,530, US 3,734,286, US 3,436,968). Characterizations of Otis US 2,919,851, Zellweger US 3,681,570 / 3,681,571, Kantarian US 3,511,372, Simjian US 3,734,286 (partial), and Bendix DE 2114676 A1 rest on (a) the '018 specification's own descriptions (which are admissions as to what those references disclose and are usable against the patentee) and (b) citation-level metadata. Every element-level mapping in § 3 must be confirmed against the full texts — column/line citations should be substituted before filing or serving any invalidity contention.
4. No litigation, IPR, PGR, or Federal Circuit activity involving US 4,792,018 was identified in the earlier searches recorded in the prior sections of this analysis, consistent with its recorded "Expired – Lifetime" status and anticipated expiration of 2005-12-20. This § 103 analysis is therefore framed as a validity challenge (e.g., defensive, or collateral to a later-patent dispute), not as an anticipation of an active enforcement proceeding. Negative search results are not proof of absence.
5. Assignee/ownership statements (Checkrobot → Omnicorp → UniQuest → NewCheck → Silicon Valley Bank → IBM) are drawn from recorded assignment data and are not legal conclusions; ownership affects standing but not the § 103 analysis.
6. Nothing above should be read as a legal opinion or as a substitute for a formal invalidity contentions chart with pin cites.

Primary sources cited

• US 4,792,018 — https://patents.google.com/patent/US4792018/en
• US 3,832,530 (Westinghouse, Object identifying apparatus) — https://patents.google.com/patent/US3832530
• US 3,734,286 (Simjian, Article handling apparatus) — https://patents.google.com/patent/US3734286 ; https://uspto.report/patent/grant/[3734286](/patent/3734286)
• US 3,436,968 (Fairbanks Morse, Processing control system) — https://patents.google.com/patent/US3436968A/en
• US 3,511,372 (Kantarian) — https://patents.google.com/patent/[US3511372A](/patent/US3511372A)
• US 2,919,851 (Otis) — https://patents.google.com/patent/[US2919851A](/patent/US2919851A)
• US 3,681,570 / US 3,681,571 (Zellweger) — https://patents.google.com/patent/[US3681570A](/patent/US3681570A) ; https://patents.google.com/patent/[US3681571A](/patent/US3681571A)
• US 4,676,343 ('913 parent) — https://patents.google.com/patent/[US4676343A](/patent/US4676343A)/en
• FR 899018 A — https://patents.google.com/patent/FR899018A/en (metadata anomaly; not relied upon)
• DE 2114676 A1 (Bendix) — https://patents.google.com/patent/DE2114676A1/en