US 4929819

I'll search for authoritative information on this patent number and any 2026 litigation activity.

Now let me search for any CAFC 2026 docket activity involving this patent.

US Patent 4,929,819 — Summary

Search coverage note: I searched for the patent number itself, for CAFC/Federal Circuit 2026 docket references, and for litigation/adverse-proceedings activity. The patent-number searches returned the authoritative record (Google Patents, USPTO.report, and third-party citations in later patents). Searches targeting a 2026 CAFC docket returned no case involving US 4,929,819 — details on that below. I did not auto-correct any identifier; everything below uses the numbers as issued.

Bibliographic data (per Google Patents record)

Source: https://patents.google.com/patent/US4929819/en

Abstract (as issued)

"An electronic module detachably mounted on a shopping cart includes a scanner for scanning bar coded information on items selected for purchase, a data processing device for data storage and control purposes and video recording apparatus capable of recording a view of the selected item as it is scanned and again as it is placed in a receptacle in the shopping cart to enable a customer to generate a record of items selected, which record can be used to facilitate checkout operations, and to provide security against improper scanning, or failure to scan, items selected for purchase."

Plain-language overview of the independent claims

Claim 1 (system). A self-service shopping system with four required cooperating pieces:

1. Multiple customer-operated carts, each with a multi-walled receptacle for goods bearing product codes.
2. A detachable module per cart, secured adjacent one end of the receptacle. Each module must contain: (a) a scanner for the product code; (b) video camera means spanning two fields of view — one covering articles being scanned, one covering articles being placed in the receptacle; (c) data processing means that associate scanned data with the recorded images and store the association; (d) a display for scanned data; and (e) a power supply.
3. Checkout means to which carts are directed, including means for receiving the module.
4. The checkout means must also total and print the stored scanned information and display the recorded video, so an operator can verify that the customer scanned and transported items correctly.

The novelty emphasis is the linked dual-view record (scan event + placement event) that is played back at checkout as a fraud/error check.

Claim 18 (method). Steps: customer selects and starts a cart → selects articles → scans each article with simultaneous video recording → places articles in the cart receptacle with simultaneous video recording → transports to checkout → cashier couples the module to data processing and video playback means → prints a transaction record from the scanned data → cashier plays back the video of the scanning and placement transactions to confirm they were properly performed → customer pays. (Claim 27 adds the step of providing an error-indication record in the module if placement is not accompanied by a scan.)

Notable dependent claims (context for scope)

• Cl. 2, 7, 8: reflective surfaces along the upper interior cart walls as a reference position, with upper/lower differently colored borders (red/blue in the spec).
• Cl. 5, 9: module activation by a motion-sensitive switch when the cart is moved (battery-saving).
• Cl. 10: optical disk drive in the module for image storage.
• Cl. 13, 14: timed recording, approximately one-fifth of a second.
• Cl. 15: error indication based on the video camera detecting an item crossing the upper colored border.
• Cl. 16, 17: error indication when an article is placed without scanner activity — via video sensing (cl. 15) or a plurality of sensors around the receptacle's upper periphery (cl. 17).
• Cl. 19, 20: module removed from the cart at checkout; battery recharged afterward.
• Cl. 22: video playback at varying speeds at the cashier's discretion.

Prior art cited on the face of the patent

US 3,836,755 (Gretag); US 3,949,194 (Plessey); US 4,071,740 (Gogulski); US 4,237,483 (Electronic Management Support); US 4,373,133 (Clyne); US 4,369,361 (Symbol Technologies); US 4,676,343 (Checkrobot); US 4,838,383 (Tokyo Electrical). The patent has 159+ "cited by" entries and also appears in the International Search Report for WO 91/20058 (Dansam Holdings), where it is cited as a "Y" reference against claims 1 and 3 (source: https://patentimages.storage.googleapis.com/78/00/03/fc86828fdd9027/WO1991020058A1.pdf).

CAFC 2026 docket check

No 2026 Court of Appeals for the Federal Circuit docket referencing US 4,929,819 was found. My searches for the patent number combined with CAFC/appeal/2026 terminology produced zero matching results; the only literal "4929819" hit outside patent sites was an unrelated Florida corporate FEI number (EDIMA Inc.), which is not an IP docket.

Confidence and caveats:

• The absence of a CAFC hit is consistent with the record: the patent expired 2008-12-12, so a live 2026 appeal on this patent would be unusual. That said, absence of a search hit is not proof of no docket — my tooling is a general web search, not a direct CAFC CM/ECF or PACER query, so I cannot rule out a non-indexed or sealed filing. If you need a definitive negative, it should be confirmed against the CAFC docket system and PACER directly.
• The "current assignee = NCR Voyix Corp" entry is a Google Patents automatic listing that carries its own disclaimer and is not an authoritative chain-of-title determination. The original assignee on the face of the patent is NCR Corporation.
• I have no reliable information about litigation history (e.g., any district court assertions) for this patent beyond the face-of-record citations; I decline to speculate on that.

I'll search for litigation involving US patent 4929819. Let me run several targeted searches.

Let me try more targeted searches on litigation databases.

Let me try several more targeted searches on litigation databases and case law.

Summary of Findings

I was unable to identify any litigation involving U.S. Patent No. 4,929,819.

This is the specific patent you asked about — "Method and apparatus for customer performed article scanning in self-service shopping," filed December 12, 1988, granted May 29, 1990, inventor Donald A. Collins, Jr., originally assigned to NCR Corporation (current listed assignee: NCR Voyix Corp).

What I searched

I ran multiple targeted queries against litigation-oriented sources, including terms combining the patent number with "litigation," "lawsuit," "case number," "infringement," "complaint," "defendant," and "plaintiff," as well as queries aimed at the Unified Patents portal, RPX Insight, and general docket sources. No case hits came back for this patent number.

What the searches did return (and why it is not your patent)

Every result mentioning "4929819" in a litigation context was a false positive for a different patent whose number merely ends in "819." Examples I explicitly set aside:

• U.S. 6,222,819 — asserted by Wi-LAN against HTC, LG, et al. (2010).
• U.S. 9,xxx,819 — various '819 patents in Apple, Treace Medical/Stryker, ASUS, and PACBIO/Oxford complaints.
• U.S. 9,899,481 — Infineon Austria v. Defendants (N.D. Cal.).
• The "FEI Number 84-4929819" appearing in a Florida corporate filing (EDIMA Inc.) — an EIN, not a patent.
• U.S. 4,929,819 appearing only as cited prior art in later patents (e.g., U.S. 5,747,784 to NCR; U.S. 5,250,789 to Johnsen; the PCT search report for WO 91/20058) — not as an asserted patent.

Assessment and caveats

• No litigation is known to me involving U.S. 4,929,819.
• Note the timeline: as a 1988-filed U.S. patent, its enforceable term ran (at most) to roughly December 12, 2008 (the "anticipated expiration" shown in the record). Any infringement suit would therefore have had to fall within roughly 1990–2008, and no such suit surfaced in my searches.
• Important limitation: My searches were run through a general web-search tool, which is not a substitute for querying the litigation databases directly. I could not confirm a complete, authoritative negative result from the Unified Patents litigation caselist, RPX Insight, PACER, or Docket Navigator for this specific number. A definitive "no litigation ever filed" conclusion would require a direct query of those platforms (and of the patent's U.S. and foreign family members, EP 0375248 A1 / JP H02-212999).

Bottom line: Based on the sources available to me, I found no known litigation — no plaintiff, defendant, jurisdiction, case number, filing date, or outcome to report — involving U.S. Patent No. 4,929,819. If you need a certified negative for a due-diligence or validity opinion, I recommend a direct PACER/Docket Navigator and Unified Patents Portal search on that exact number, since I cannot rule out an unfiled, sealed, or non-digitized older case that my search tools did not index.

Proceedings overview

Total AIA trial proceedings on file for US 4,929,819: 0 — no IPRs, no PGRs, no CBM reviews, in any status (nothing instituted, nothing invalidated, nothing settled, nothing denied). The USPTO ODP structured block is the canonical source here, and my independent web check surfaced no IPR/PGR/CBM petition, no Board Final Written Decision, and no Federal Circuit appeal involving this patent. Bottom line for a defendant: this is not "a patent hardened by surviving IPRs" — it is a 1990-issued, long-expired patent with no PTAB history at all, which means your defensive leverage lies entirely outside the AIA trial regime (expiry, § 286 time bar, and § 102/§ 103 prior art if anyone resurrects it), not in an estoppel-laden IPR record.

---

No proceedings to report

There is no proceeding number to populate for the per-proceeding template. I did not invent one, and I will not. For completeness, the per-proceeding fields resolve as follows:

• Type — N/A (no AIA trial on file)
• Filed — N/A
• Status — N/A. Note the patent's own legal status per the source data: "Expired - Lifetime", with an anticipated expiration date of 2008-12-12 (20 years from the 1988-12-12 filing).
• Judge panel — N/A (no panel ever assembled)
• Petition grounds — N/A
• Institution decision — N/A
• Final Written Decision — N/A. No claim of US 4,929,819 has ever been canceled, confirmed, or otherwise adjudicated at claim level by the Board. All 27 claims stand as issued, for whatever that is worth on an expired patent.
• Settlement / termination — N/A
• Appeal — N/A. No CAFC docket number exists, because there is no underlying Board decision and I found no district court judgment to appeal.
• Defensive value — See below; the value is real but comes from the calendar, not from a PTAB record.

---

False-positive check (important — do this before you rely on my "zero")

PTAB is an acronym-dense place, and web queries for "4929819" / "'492 patent" are noisy. My search returned two ptacts.uspto.gov Final Written Decision excerpts (petition artifacts 1542007 and 1534084) that styling-identify as "the '492 patent" — these are not US 4,929,819. Their claim sets (e.g., claims 1–5, 9–16, 20–27, 31–33) and subject matter ("source systems" evaluated against "target systems" using "at least one rule set" — i.e., data-migration/placement technology) have no relationship to the 27 claims of 4,929,819, which are directed to a cart-mounted scanner module plus video-recording module with reflective cart walls. Anyone pulling those FWDs into a memo as "the IPR on the NCR self-scan patent" is citing the wrong patent family. To confirm zero before filing anything, run the patent number directly (not the "'492" shorthand) through PTAB E2E / Patent Trial and Appeal Board Center and through Patent Public Search's "PTAB" and "Litigation" tabs.

One corroborating signal from my search: 4,929,819 shows up in the literature only as cited background prior art — e.g., it is cited in CN104299336B and in the specification of US 20140316901A1 (limited-time merchandise promotion) as an example of a conventional "smart shopping cart." A patent that is only ever cited as prior art and never challenged at the Board is behaving exactly like an expired reference, not like an assertion asset. I also found no district court complaint asserting this patent; the NCR enforcement activity that surfaced in my search (e.g., a 2014 D. Del. complaint by NCR via Kilpatrick Townsend) involved different NCR patents, not '819. I state that as "found no evidence of," not as a verified litigation-clearance result.

---

Strategic summary

Claim status. There is no CANCELED / SUSTAINED breakdown to give you, because no tribunal has ever touched these claims. All 27 claims — independent system claim 1, independent method claim 18, and their dependents — are formally UNTESTED by the PTAB. The more useful framing for a defendant is that the distinction is academic: the patent's statutory term ran out on 2008-12-12, roughly four years before AIA trials became available (IPR/PGR/CBM practice began 2012-09-16). That timing, not the absence of a disgruntled competitor, is the straightforward explanation for a null PTAB docket on a 1988-filed NCR patent. The "well-asserted patents eventually attract IPRs" heuristic does not apply to a patent that expired before the forum existed.

Estoppel landscape — this is the good news, and it cuts your way. Because no IPR/PGR was ever instituted, § 315(e)(2) estoppel is a non-issue: no petitioner or privy is barred, and there is no Board record to be trapped by. Practically, though, the estoppel question is downstream of two harder defenses. First, § 286: an expired patent can only reach pre-expiration infringement, and damages are recoverable for at most the six years preceding the filing of the complaint. Since expiration was 2008-12-12, any suit seeking damages would have had to be filed by roughly 2014-12-12; as of today (2026-10-01) that window is long shut, and no live infringement theory exists for post-expiration conduct because an expired patent cannot be infringed. Second, if you are nevertheless facing a demand letter built on this patent, the demand is essentially an invalidity/prior-art bluff, and you are free to use the whole field of § 102/§ 103 art without any estoppel constraint — including the very prior art already of record: US 3,836,755 (Gretag), US 3,949,194 (Plessey), US 4,071,740 (Gogulski, "Mobile automated shopping system"), US 4,237,483 (Electronic Management Support), US 4,369,361 (Symbol), US 4,373,133 (Clyne), US 4,676,343 (Checkrobot), and US 4,838,383 (Tokyo Electrical) — the eight references the examiner cited, all of which predate the 1988-12-12 filing.

Pattern signals. No repeat petitioner (there is no petitioner). No patent-owner PTAB appeals (there is no Board decision to appeal). No defensive aggregator in the chain — I found no Unified Patents, RPX, or similar filing. The assignee chain is unremarkable: filed by NCR Corp., original assignee NCR Corporation (a Corp. of MD), current assignee listed as NCR Voyix Corp.; this is ordinary corporate history rather than a troll assignment, which is consistent with the patent being a legacy portfolio artifact rather than an enforcement vehicle.

---

Recommended next steps

1. Treat this as a zero-proceeding patent, and document the check. Pull the patent number (not "'492") through PTAB E2E / PTAB Center and Patent Public Search's PTAB and Litigation tabs, and screenshot the null result. That contemporaneous record is your rebuttal if opposing counsel asserts "this patent survived IPR" — it never faced one.
2. Lead with expiry and § 286, not with PTAB. If a demand letter cites US 4,929,819, the dispositive response is that the patent expired 2008-12-12 (anticipated expiration per the source record; the patent's own legal status is "Expired - Lifetime") and that any pre-expiration damages claim is time-barred under 35 U.S.C. § 286. There is no FWD to link to, so quote the source record's expiration entry and the "Expired - Lifetime" status instead.
3. Keep the record art staged, not briefed. Because no estoppel attaches, you retain the ability to run § 102/§ 103 against any surviving claim should someone assert pre-expiration conduct in a timely-related action. US 4,071,740 (Gogulski) — a mobile automated shopping system with on-cart item handling — is the reference to read first against claims 1 and 18.
4. Do not cite the ptacts.uspto.gov "'492 patent" FWDs. Those decisions concern a different, data-migration '492 patent. I could not verify their patent number within my search budget and I am not asserting one; the claim sets and subject matter alone establish they are unrelated. If you need their number for a separate matter, retrieve it directly from the cited petition artifacts (1542007, 1534084) rather than inferring it.
5. One honest caveat. My search effort was bounded, and I could not exhaustively run every litigation docket (PACER, Docket Navigator, CourtListener RECAP) for this patent number. I found no evidence of district court assertion or CAFC appeal, but "no evidence found" is weaker than "confirmed none." If a demand letter is actually in hand, that is the one gap worth closing with a paid docket search before you rely on the expired-patent posture in a responsive brief.

I'll pull the authoritative record and cross-check the assignment chain before writing this up.

US 4929819 — Ownership Chain Reconstruction

Data-access caveat up front: I could not open assignmentcenter.uspto.gov or assignment.uspto.gov directly in this session (no direct-fetch tool; search budget exhausted after the queries below). The chain below is built from the authoritative patent text you supplied (Google Patents legal-events block) plus corroborating open sources. Two fields I could not retrieve and will not guess: the reel/frame of the 1988 recording, and the correspondent of record on it. I flag every place this matters rather than inventing a number.

Cross-checked sources: Google Patents US4929819A, plainsite assignment 69419/561, NCR name-change press coverage, NCR Voyix/Atleos cross-license disclosure.

---

Inventors

Unusual-pattern check: none observed. Single inventor, no co-inventor mismatch, no evidence of inventor departure within 12 months of the 1988-12-12 filing, and no inventor-originated later assignment. A one-inventor, one-employer chain with no inventor-side activity is the opposite of the pre-fire-sale signature (which requires inventors departing or individually retaining/reassigning rights). I cannot verify Collins's tenure at NCR beyond the filing date — that is a genuine gap, not a negative finding.

---

Original assignee

NCR Corporation, a Maryland corporation (principal address at the time of the specification: Dayton, Ohio; corporate headquarters later relocated to Duluth/Atlanta, Georgia). Named as assignee on the face of the issued patent. Current owner of record per Google Patents' legal-events block: NCR Voyix Corp (see below).

• Primary line of business: retail/commerce technology — point-of-sale terminals, self-checkout (SCO) kiosks, ATMs, and retail software. NCR was the incumbent supplier of the checkout hardware this patent builds on. It is unambiguously an operating company, not a licensing vehicle.
• Did they ship a product embodying the claims? Partly verifiable, partly not. The claims cover a cart-mounted, detachable scanner module with dual-field-of-view video audit and a checkout-side playback terminal — an aggressive 1988 architecture. I have no confirmed evidence NCR commercialized this exact cart-module configuration. What is confirmed is that NCR kept filing in the same technical lane off this disclosure's family: the spec/EP family includes EP0847032 ("Enhanced security self-service checkout station," applicant NCR International, Inc., Dayton, Ohio) and later NCR filings US5747784 (enhancing security in a self-service checkout station) and USRE41093 (monitoring item shuffling in a post-scan area of a self-service checkout terminal) both cite this patent. That is a strong internal-continuation signal but not proof of an embodying product. Mark: unclear on embodiment; not present on abandonment.
• Current status: Operating. NCR Corporation changed its name to NCR Voyix Corporation effective 2023-10-13, concurrent with the spin-off of NCR Atleos Corporation (closed 2023-10-16); NCR Voyix trades on NYSE under VYX. Voyix and Atleos entered a Patent and Technology Cross-License Agreement dated 2023-10-16 covering each party's patents and technology. No bankruptcy, no dissolution. Note the legal status field on the patent itself is "Expired – Lifetime" with anticipated expiration 2008-12-12 — the full 20-year term from the 1988-12-12 filing, i.e. expiry by term, not by maintenance-fee lapse or sale.

---

Assignment timeline

The Google Patents legal-events block shows exactly one operative recorded transfer, executed and recorded on the filing date:

• 1988-12-12 (executed) / recorded 1988-12-12 — Reel/Frame not captured in this session (Assignment Center not directly queryable; Google Patents does not expose reel/frame for this record)
  • Conveyance: Assignment (cap. "ASSIGNMENT OF ASSIGNORS INTEREST")
  • Assignor: Donald A. Collins, Jr.
  • Assignee: NCR Corporation, a Corp. of MD
  • Correspondent: not determined — could not retrieve from the indexed record. I will not fabricate a name or firm here.
  • Context: Standard employee-invention assignment to the applicant, executed on the filing date. Not a fire-sale, not a reorg, not a securitization.

One further entry, flagged as not confirmed against this patent number:

• 2023-10-13 (executed) / recorded 2024-11-27 — Reel/Frame 69419/0561 (plainsite renders it "Patent Assignment 69419/561")
  • Conveyance: Change of Name
  • Assignor: NCR Corporation → Assignee: NCR Voyix Corporation
  • Correspondent: Schwegman, Lundberg & Woessner, PA, P.O. Box 2938, Minneapolis, MN 55402
  • Context: Internal corporate name change following the NCR Atleos spin-off. This is the recording that explains Google's "Current Assignee: NCR Voyix Corp" line.
  • ⚠️ Confidence limit: the record I located is indexed to a different NCR property ("Safe Door Apparatus And System," app. 18/937,656). USPTO change-of-name recordings are frequently filed as large multi-property batches, and this is consistent with a blanket NCR→Voyix chain — but I have not verified that US 4,929,819 appears on reel 69419/0561. Treat as inferred, not confirmed. Note also that on its face this is a purely nominal change: any NPE analysis is unaffected, and the patent had already been expired ~15 years when executed.

No security agreements, no mergers, no releases, no corrections, no assignments to any non-operating entity appear anywhere in the record.

---

Timeline diagram

timeline
    title Ownership of US 4929819
    1988 : Filed by NCR Corporation
         : Inventor Collins assigns rights to NCR
    1990 : Patent issues as US 4929819
    2008 : Term expires 20 years from filing
    2023 : NCR Corporation renamed NCR Voyix

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The only recorded conveyance runs to NCR Corporation (reel/frame not captured, 1988-12-12), an operating manufacturer named on the patent face. There is no downstream transfer to any "IP / Holdings / Licensing / Ventures" entity anywhere in the record. Critically, the patent's expiration on 2008-12-12 post-dates every recorded transfer, so there was never a live right to move.

2. Known asserter in the chain — not present. Contrast against the named lists (Acacia, Marathon, IV, IPNav, Wi-LAN/Conversant, Vringo, Pendrell, Round Rock, etc.): no such entity appears as assignee, assignor, or correspondent. The downstream citers of this patent include asserter-adjacent filers, but citation is not ownership and I treat it as noise. The current owner of record, NCR Voyix, is a NYSE-listed operating company with no NPE profile.

3. Repeat correspondent across the chain — unclear, and I want to be precise about why. With only one operative link I have a chain of length 1, which structurally cannot exhibit recurrence. Moreover I could not retrieve the 1988 correspondent of record at all, so I have no name to test for recurrence. The only correspondent name I surfaced (Schwegman, Lundberg & Woessner on the 2023 name change, reel 69419/0561) is a large general-practice IP firm that does routine prosecution and corporate housekeeping for operating companies — a single appearance, and on a name-change filing, is explicitly not a finding under your own rule. No flag.

4. Cascading transfers — not present. Zero chained-LLC transfers; no two assignees share a correspondent address in this record because there is only one assignee of substance.

5. Pre-litigation transfer — not present. I found no infringement suit naming US 4,929,819. The single recorded assignment predates issuance by ~17 months and is a payroll/employment assignment to the applicant, which is the archetype of an ordinary rather than an arranged-for-venue transfer. (Residual limit: I did not run a full Docket Navigator / PACER sweep; litigation absence is a negative finding from my searches, not a certified null.)

6. Bankruptcy fire-sale — not present. NCR Corporation has no Chapter 7/11 history; the 2023 event was a voluntary two-way spin-off (NCR Voyix / NCR Atleos) that added a cross-license rather than liquidating a portfolio.

7. Privateering — not present. No transfer of this patent to any third-party NPE, and no SEC or trade-press report of NCR asserting this patent through an intermediary. The 2023-10-16 Voyix/Atleos Patent and Technology Cross-License Agreement is intra-group licensing between two NCR successor operating companies — the opposite of privateering.

8. Defensive aggregator (anti-NPE) — not present, but functionally achieved. The chain does not terminate at RPX, AST, LOT, Unified, or OIN. It terminates at NCR Voyix, which is an operating company. Separately, the patent is expired, so no aggregator would acquire it — neutralization happened by operation of law on 2008-12-12 rather than by purchase.

---

Verdict

Insufficient data — in the sense your rubric defines it: only the original assignment exists, and no NPE thesis can be built on it.

Justification: the record contains a single operative conveyance — Collins → NCR Corporation, a Corp. of MD, executed and recorded 1988-12-12 — with no reel/frame or correspondent recoverable in this session, and no security interest, merger, or downstream transfer of any kind. The only other candidate entry, a Change of Name from NCR Corporation to NCR Voyix Corporation (executed 2023-10-13, recorded 2024-11-27, reel 69419/0561, correspondent Schwegman Lundberg & Woessner), is inferred rather than confirmed to list this patent number, and is purely nominal in any event. All eight NPE signals are negative or unclear, and the patent expired 2008-12-12 on the full 20-year term, so assertion was moot ~15 years before the most recent ownership event. The affirmative complementary finding is that US 4,929,819 was held continuously by an operating company (NCR → NCR Voyix) for its entire enforceable life — there is no shell-entity link, no known asserter, no repeat NPE correspondent, and no litigation.

Verify directly (the reel/frame gap is the one thing worth closing):

• Assignment Center, search by patent number 4,929,819: https://assignmentcenter.uspto.gov/
• Legacy Assignment Search mirror: https://assignment.uspto.gov/patent/index.html
• Google Patents legal events: https://patents.google.com/patent/US4929819/en

What to look for when you run it: the reel/frame and correspondent of record on the 1988-12-12 Collins→NCR entry, and whether reel 69419/0561's property list actually includes 4,929,819. If the only two hits are those, the "Insufficient data / clean operating-company chain" call is final. If a third entry exists that my sources did not surface, it would have to post-date 2008-12-12 to involve a live right.

I'll search the USPTO/patent databases for US 4,929,819 specifically, then work through each of its cited references.

Good coverage so far. Now the remaining five cited references.

Prior-Art Analysis — US 4,929,819 (Collins / NCR)

1. Target patent confirmed (not a look‑alike number)

I retrieved the record for US 4,929,819 only — no results for other numbers ending in "819" are reported below.

Sources: https://patents.google.com/patent/US4929819/en ; https://uspto.report/patent/grant/[4929819](/patent/4929819) ; full‑text PDF https://patentimages.storage.googleapis.com/8b/81/70/928fdfc750281b/US4929819.pdf

Analysis date note: The task states today is April 26, 2026. The patent has been expired since 2008‑12‑12, which matters for the practical relevance of any §102 analysis below.

---

2. Method, scope, and honesty flags

• The eight references below are the "Patent Citations (8)" of record for US 4,929,819, and each carries the Google Patents asterisk (*) = "cited by examiner" marker. So all eight were examiner‑cited, not third‑party‑cited.
• I verified the bibliographic data (numbers, dates, titles, assignees) directly against the patent record and, where retrieved, against the primary documents.
• I fully retrieved and read the claims/disclosure substance of four references this session: US 3,836,755, US 3,949,194, US 4,071,740, US 4,238,483.
• ⚠️ I was not able to retrieve the disclosure text for US 4,373,133, US 4,369,361, US 4,676,343, US 4,838,383 before my search budget was exhausted. For those four I give the verified citation/dates from the record, and I describe their subject matter from general knowledge clearly labelled as unverified pending a full‑text check. Do not treat those descriptions as confirmed quotes.
• Legal standard caution: A dependent claim incorporates every limitation of the claim from which it depends. Since all 27 claims depend from independent claim 1 or claim 18, no reference can anticipate a dependent claim unless it also discloses all elements of claim 1 (or 18). This drives the conclusion in §4.

---

3. The eight references of record — element‑level analysis

3.1 US 4,071,740 — Gogulski — most relevant reference

• Full citation: US 4,071,740 A, "Mobile automated shopping system," Paul Gogulski. Filed/priority 1976‑05‑26; issued 1978‑01‑31. Pre‑AIA §102(b) art (issued >1 yr before 1988‑12‑12).
• Disclosure (verified): A shopping cart with a control panel carrying a hand‑held scanner, a second stationary scanner on the panel, a paper‑tape printer, and multiple visual displays (product description, cost, unit price, sale). A weight‑sensing device in the cart's product‑holding area and a light‑sensing system on the cart's top perimeter detect entries by "breaking the light plane." An unauthorized entry (item placed without a scan) lights a reject light and can stop the cart. Prices are fetched from an in‑store computer; a "dual scanning cross‑check" is described.
• Claim mapping:
  • Cart + on‑cart scanner + display + printer architecture → concept of claim 1, claim 12 (display of item description and price each time an article is scanned).
  • "Plurality of sensing devices mounted around the upper periphery of the cart receptacle" producing an error/reject indication when an item is placed without scanning → claim 17, and the function of claim 16.
  • "Travel button light indicating the cart is in movement" → relates to the concept behind claims 5/9 (module activation tied to cart movement).
• §102 assessment: Does not anticipate claim 1 or claim 18. It lacks the dual‑view video camera means (scan area + receptacle‑placement area), the detachable module with associated image/data storage, and the checkout‑side video playback for operator verification. It is the strongest individual §103 reference against claims 1/16/17/12, and the single best piece of art for the "cart‑top sensing + error indication" dependent claims.

3.2 US 3,836,755 — Gretag AG

• Full citation: US 3,836,755 A, "Self‑service shop," Gretag AG (Switzerland). Filed/priority 1972‑02‑14; issued 1974‑09‑17. §102(b) art.
• Disclosure (verified): Purchase trolleys each with a container; a reading device for the machine‑readable identifying data of each article placed in or removed from the container; weighing means correlated with the container. A first electronic circuit signals when the data was read correctly but no article was placed/removed, and when an article was placed/removed but the data was not read correctly. Includes an intermediate weight store, comparator, error‑recognition device, alarm stages, and a connector on the trolley that couples the trolley stores/comparator to a check‑out station computer (or radio transmission).
• Claim mapping:
  • Error indication when an article is placed in the receptacle without a corresponding (correct) read, and an error store at the check‑out station → the function of claim 16 and the error‑record aspects of the claim 18/27 method.
  • Trolley‑to‑check‑out‑station coupling of stored data for the checkout transaction → concept of claim 1(h)/(i) and claim 18 ("coupling said module to the data processing means").
• §102 assessment: Does not anticipate claim 1 or claim 18 — it uses weight comparison, not video, and has no dual‑view camera, no display‑of‑scanned‑data element as claimed, and no video playback. It is a strong §103 reference on the "unscanned‑item error" dependent claims and evidence that cart‑side detection + checkout data transfer was known.

3.3 US 4,238,483 — Electronic Management Support, Inc. (Clever)

• Full citation: US 4,238,483 A, "Surveillance system," inventor Eric C. Clever, applicant Electronic Management Support, Inc. Application filed 1979‑03‑19; earliest priority 1976‑12‑22; issued 1980‑12‑02. §102(b) art.
• Disclosure (verified): A CCTV camera (or cameras) used in conjunction with electronic cash registers at point‑of‑sale stations. Combined images of the transaction items (from the camera) and the transaction data rung up on the register are presented on a TV monitor; the composite video signals are stored on video tape together with coded signals related to transaction data for computerized editing. Includes character generation, date/time data, multi‑camera switching.
• Claim mapping:
  • Associating scanned/registered transaction data with recorded video, and storing the association → the core of claim 1(e) ("data processing means for associating the information scanned by the scanner with the images recorded by the video camera means and for storing such associated information").
  • Display of a video record to an operator to verify a transaction → the function of claim 1(j) and the playback step of claim 18.
• §102 assessment: Does not anticipate claim 1 or claim 18 — it fixes the camera at the POS station, not on the cart; there is no cart‑mounted detachable module, no on‑cart scanner, no dual‑view scan/placement record, and no separate display of scanned data to the customer. It is nonetheless the most on‑point §103 reference for the "video + transaction data are stored together and reviewed by an operator" concept.

3.4 US 4,676,343 — Checkrobot Inc. (Humble et al.)

• Full citation: US 4,676,343 A, "Self‑service distribution system," Checkrobot Inc. Filed 1984‑07‑09; issued 1987‑06‑30. §102(b) art (issued >1 year before 1988‑12‑12).
• Disclosure: Unverified this session (full text not retrieved). Based on the title/assignee and general knowledge, this is the well‑known Checkrobot self‑service supermarket checkout art in which the customer scans items and a security/verification mechanism detects items that were not properly scanned. To be treated as provisional until the full text is checked.
• Claim mapping (provisional): Self‑service scanning + anti‑shrinkage verification → context for claim 1 and claim 16.
• §102 assessment: No anticipation of claim 1/18 on the face of it (no cart‑mounted dual‑view video module, no image‑to‑scan association, no checkout video playback). Relevant as §103 background art and possibly as evidence of the general self‑service/security goal.

3.5 US 4,373,133 — Nicholas Clyne

• Full citation: US 4,373,133 A, "Method for producing a bill, apparatus for collecting items, and a self‑service shop," Nicholas Clyne. Filed 1980‑01‑03; issued 1983‑02‑08. §102(b) art. Note the record's "Family Cites Families" list also includes GB 2068132 B (Wright L.H., priority 1980‑01‑03, published 1984‑03‑14), "Method for producing a bill during the collection of items in a self‑service shop," which appears to be the GB counterpart of the same subject matter.
• Disclosure: Unverified this session. Per title and general knowledge: self‑service shop in which items are collected and a bill/tally is produced for the customer's collected items (i.e., pre‑checkout recording of selected items).
• Claim mapping (provisional): Producing a purchase record from items collected during shopping → peripheral to claim 1(i) and the printing step of claim 18.
• §102 assessment: No anticipation of any claim on the record before me; likely §103 background only.

3.6 US 4,838,383 — Tokyo Electrical Co., Ltd. — §102(e)‑only reference

• Full citation: US 4,838,383 A, "Apparatus for checking out purchases," Tokyo Electrical Co., Ltd. Filed 1987‑03‑16; issued 1989‑06‑13.
• ⚠️ Critical timing point: Its issue date (1989‑06‑13) is after the 1988‑12‑12 filing/priority of US 4,929,819. It therefore cannot be §102(a) or §102(b) art; it is available at most as pre‑AIA §102(e) art (U.S. patent granted on an application filed 1987‑03‑16, before the '819 filing).
• Disclosure: Unverified this session (title indicates a purchase‑checkout apparatus).
• §102 assessment: Cannot anticipate on §102(a)/(b) grounds. Only §102(e) potential, and only for whatever it discloses — which, on the record before me, does not include the claimed dual‑view cart‑mounted module/checkout‑video combination.

3.7 US 4,369,361 — Symbol Technologies

• Full citation: US 4,369,361 A, "Portable, stand‑alone, desk‑top laser scanning workstation for intelligent data acquisition terminal and method of scanning," Symbol Technologies, Inc. Filed 1980‑03‑25; issued 1983‑01‑18. §102(b) art.
• Disclosure: Unverified this session. Per title: a laser scanning workstation (scanner hardware/reading method).
• Claim mapping (provisional): Scanner element of claim 1(c) (and the scanner in claim 18) — but as a hardware teaching, not the claimed combination.
• §102 assessment: No anticipation of claim 1/18; relevant only as §103 art showing the scanner element itself was old.

3.8 US 3,949,194 — Plessey Handel Und Investments A.G.

• Full citation: US 3,949,194 A, "Check‑out terminal," Plessey Handel Und Investments A.G. Filed/priority 1973‑04‑04; issued 1976‑04‑06. §102(b) art.
• Disclosure (verified): A conveyor system for a checkout terminal — a tilted main belt plus subsidiary belts, with one or more scanning slits and an optical code reader detecting coded labels on the side of articles as they pass; several belt/roller arrangements for single‑ and multi‑slit scanners; a checkout terminal incorporating any of these conveyor systems.
• Claim mapping: At most the concept of scanning product codes at a checkout station (claim 1(c)/(h) context).
• §102 assessment: Least relevant of the eight. No cart, no module, no video. No anticipation of any claim; marginal §103 art.

---

4. Bottom line on §102 vs. §103

No single one of the eight cited references anticipates independent claim 1 or independent claim 18. Each is missing at least one of the following limitations that the independent claims require:

1. a plurality of detachable modules, one per cart, secured to the cart (claim 1(b));
2. video camera means covering two distinct areas — the scan area and the receptacle‑placement area — in one module (claim 1(d); method step of claim 18);
3. data processing means that associate the scanned data with the recorded images and store the association (claim 1(e)) — only US 4,238,483 comes close, and it does so with a fixed POS camera, not a cart module;
4. checkout means that receive the module and display its recorded video so an operator can verify the customer's scan/placement actions (claim 1(h)–(j); claim 18 playback step) — no cited reference discloses cart‑captured dual‑view video played back at checkout.

Because every one of claims 2–27 depends from claim 1 or claim 18, none of the 27 claims is anticipated by any of these eight references taken alone. Their correct role is:

• §103 combination candidates (primary): US 4,071,740 (Gogulski) + US 3,836,755 (Gretag) + US 4,238,483 (Clever) — cart‑mounted self‑scanning with sensing/error indication, plus video‑with‑transaction‑data recording.
• Feature‑only / secondary §103 art: US 4,676,343 (self‑service checkout security), US 4,373,133 (self‑service billing), US 4,838,383 (§102(e) only), US 4,369,361 (scanner hardware), US 3,949,194 (checkout conveyor scanning).

If you want a per‑claim "best single reference" list for an obviousness chart, the tightest mappings are:

---

5. References the record suggests you should also check (not in the eight)

These appear in the record but are not among the eight US citations examined above; two are potentially material §102 art because they predate 1988‑12‑12:

• US 4,779,706 A — NCR Corporation, "Self‑service system," filed 1986‑12‑17, issued 1988‑10‑25. Listed under "Family Cites Families." NCR's own earlier self‑service art. Issued ~7 weeks before the '819 filing — available as §102(a) art, and as §102(e) art via its 1986‑12‑17 filing date. This is the single most important additional reference to verify, given the shared assignee and subject matter.
• US 4,630,110 A — Supervision Control Systems, Inc., "Surveillance system," filed 1984‑02‑15, issued 1986‑12‑16. Also in "Family Cites Families." §102(b)‑eligible video‑surveillance art.
• US 3,749,191 A (R. Ashen) and GB 2 068 132 B (Wright L.H.) — additional family citations, §102(b)‑eligible; GB 2 068 132 appears to correspond to the US 4,373,133 / Clyne subject matter.
• WO 91/20058 A1 (Dansam Holdings) — ⚠️ not prior art here: it was published 1991‑12‑26, after the '819 filing, and it cites US 4,929,819 as a "Y" reference against its own claims 1 and 3. It is a forward citation, not a §102 reference against '819.

---

6. Caveats and what I could not confirm

1. Four references described only provisionally — US 4,373,133, US 4,369,361, US 4,676,343 and US 4,838,383 disclosures were not retrieved in this session (search budget exhausted). Their bibliographic data is verified from the record; their subject‑matter descriptions are from general knowledge and must be confirmed against full text before being relied on in any §102/§103 chart.
2. Examiner vs. third‑party citation — all eight carry the examiner (*) marker in the retrieved rendering; I could not independently confirm this against the printed front page of the patent.
3. US 4,238,483 date nuance — the issued patent's application was filed 1979‑03‑19 with earliest priority 1976‑12‑22 (Espacenet). Either date makes it §102(b) art; I note both for accuracy.
4. Identifiers left literal. I did not auto‑correct anything. In particular, a Board of Patent Appeals decision (Ex Parte Shimizu, Appeal No. 1999‑1625, App. 08/792,468) discusses a §103 rejection over "Collins and Gogulski," with the appellant referring to "Collins, Jr. '947." I flag this literally: I cannot confirm that "Collins '947" is US 4,929,819, and the notation does not match this patent's number. It is not asserted here as a citation to '819.
5. No 2026 docket activity for US 4,929,819 was identified in the prior section of this analysis, consistent with the 2008‑12‑12 expiry.

I'll verify the actual disclosures of the prior-art references of record before mapping them to the claims.

Obviousness Analysis — US 4,929,819 under 35 U.S.C. § 103

Subject patent: US 4,929,819 A, "Method and apparatus for customer performed article scanning in self-service shopping" (Collins, Jr.; NCR Corp.), App. 07/283,116, filed 1988‑12‑12, issued 1990‑05‑29, 27 claims.

Scope note / methodology. You directed me to the Prior Art section of the US4929819 page. I worked from the Patent Citations (8), Family Cites Families (4), and the Cited By listings, and I independently pulled the full text/claims of four of the cited references (Gretag/Ehrat, Gogulski, Clyne, Checkrobot) plus partial text of US 4,838,383. I hit my tool-step limit before verifying the remaining four of the eight (Plessey, Electronic Management Support, Symbol, Tokyo Electric) and the four family cites. Those are characterized below from the face of the record (title/assignee/date) and my background knowledge, and are flagged as unverified — any contention-grade invalidity chart should confirm their disclosures against the USPTO Patent Public Search full texts. I have not auto-corrected any number.

Cross-reference flags against the prior section:

1. The earlier section did not note that US 4,838,383 (Tokyo Electric) issued 1989‑06‑13 — after the 1988‑12‑12 filing date of '819. It therefore cannot be §102(a)/(b) art; it is available only as §102(e) art via its 1987‑03‑16 US filing.
2. The earlier section treats all eight citations as undifferentiated "prior art." They are not coequal — see the status column in the table below. This matters to the combination theory.
3. Newly surfaced and relevant: the Board of Patent Appeals decision Ex parte Shimizu, Appeal No. 1999‑1625, App. 08/792,468, sustained a §103 rejection "unpatentable over Collins and Gogulski" for cart-attached scanning subject matter. The reference is designated "Collins, Jr. '947" (i.e., US 5,149,947, Sept. 22, 1992 — consistent with the justia citation listing "5149947 … Collins, Jr."), not '819; the column/line citation given in the decision does not correspond to '819. I flag this so it is not mis-cited as the '819 patent being invalidated. Source: https://law.onecle.com/board-of-patent-appeals/2002/fd991625010.html

---

1. Governing law and level of ordinary skill

The application was filed before 2011‑09‑16, so pre‑AIA § 103(a) governs. The framework is Graham v. John Deere, 383 U.S. 1 (1966), as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) and MPEP § 2143. The threshold inquiry is whether the differences between the claim and the prior art are such that the subject matter as a whole would have been obvious to a POSITA at the time of invention (December 1988), considering (i) scope/content of the art, (ii) differences, (iii) skill level, and (iv) objective indicia.

Hypothetical POSITA (Dec. 1988): a person with a B.S. in electrical engineering or computer science (or equivalent) and about 2–3 years of experience in point-of-sale/retail automation, or a technician of comparable practical experience, familiar with: UPC bar-code scanning (slot scanners, laser scanners), portable data terminals and battery-powered mobile electronics, CCTV/video surveillance recording and playback, and cart-based data-collection systems. This framing is favorable to the patentee — it is a relatively high-skill artisan who would readily design the claimed combinations.

Claim-construction notes that drive the analysis:

• "video camera means capable of covering a first area … and a second area" (cl. 1) is means‑plus‑function language (pre‑AIA §112 ¶6). Under the '819 specification the disclosed structure is a single camera 45 whose field of view is split by a single-sided mirror 47 (FIGS. 5C, 7). But the claim covers the corresponding structure and equivalents, so a two-camera or mirror/splitter implementation satisfies the limitation. This substantially weakens any argument resting on the specific single-camera/split-field structure.
• "associating the information scanned by the scanner with the images" (cl. 1) is a data-processing function with no specific algorithm recited. Time-indexing or transaction-flagging a video frame to a scan event is routine programming.
• "detachable module," "secured to said carts adjacent one end of said receptacle" – structural placement/attachment; the art is rich in cart-mounted and cart-attachable electronics.

---

2. The prior art of record (status and disclosure)

"Cited By" entries (e.g., US 5,412,193 Symbol mobile POS; US 5,250,789 Johnsen; US 5,340,671‑family; US 5,729,697/6,032,127 IBM intelligent shopping cart) all post-date '819 and cannot be §103 art against it. They are useful only as corroboration that the cart-mounted scanner + on-cart data processing/X-Y verification architecture was the recognized technical trajectory — not as references in a rejection.

---

3. Claim 1 (system) — the primary combination

Proposed rejection: Ehrat (US 3,836,755) in view of Gogulski (US 4,071,740), further in view of a surveillance-recording reference (US 4,237,483 and/or US 4,630,110), and optionally Clyne (US 4,373,133).

Rationale for why the combination is not merely hindsight: every element was known for its established function, the art addressed the identical problem (unscanned/unrecorded merchandise entering the cart), and the references themselves expressly motivate the substitution — see §7.

Alternate primary combinations (for robustness)

• Combination B (cart-centric): Clyne + Gogulski + US 4,630,110. Clyne supplies the cart-mounted recording unit, checkout data transfer, print-out, and the express "challenge a dishonest customer" purpose; Gogulski supplies the on-cart scanner, display, printer, removable bag, rechargeable battery, and top-perimeter sensing; the surveillance reference supplies the recording camera and later playback.
• Combination C (checkout-centric): Checkrobot + Gogulski + US 4,237,483. If a tribunal reads claim 1's module as requiring only cart-attachable scanning/recording (not a full cart-based system), Checkrobot's self-scan + security-zone + rejection-on-unverified-entry teaching plus Gogulski's cart hardware and a surveillance camera completes the claim.
• KSR "obvious to try" fallback: as of 1988 the art identified a finite set of known verification mechanisms for the unscanned-item problem — weight comparison (Gretag, Clyne, Checkrobot), light-plane/optical curtain (Gogulski, Checkrobot), and video surveillance (US 4,237,483, US 4,630,110). Selecting and implementing one of these known options with a reasonable expectation of success is obvious under KSR.

---

4. Claim 18 (method) — parallel combination

Proposed rejection: Ehrat in view of Gogulski, in view of Clyne and US 4,630,110 (or US 4,237,483).

Note the method claim is easier to defeat than claim 1, because it does not require the dual-field-of-view camera or the association/storage function — only "video recording means" recording the scan and the placement. Two cameras, or a single recording pass covering the cart front, satisfy it.

---

5. Dependent claims

Group A — reflective surfaces / colored borders (cl. 2, 7, 8, 23, 24, 25). Gretag/Gogulski establish cart-mounted reference/trip mechanisms. Gogulski's light-sensing plane at the cart's top perimeter and Checkrobot's optical curtain perform the same function (defining a plane whose breach signals an unscanned placement). Substituting mirrors with differently colored upper/lower borders as the reference fiducials for a camera's view is the kind of substitution of one known sensing reference for another that KSR holds obvious; the red/blue stripe selection is an unrebutted design choice. This is the weakest part of any rejection — no verified reference discloses mirrors with two differently colored borders, and the patentee can argue the specific fiducial geometry is not suggested. Expect this to be the battle line, and expect it to be resolved as "predictable variation" unless the patentee proves unexpected results.

Group B — detachable module / removal at checkout / recharge (cl. 3, 4, 19, 20). Ehrat cl. 9 (connector to checkout), Clyne (unit taken by assistant), Gogulski (rechargeable battery, recharged when not in use). Checkrobot's criticism of Gretag's per-cart electronics ("a large supermarket … might require 50 to 100 or more such carts … can not be nested … major storage problem") supplies a strong express motivation for a module that can be shared/moved among carts and rack-charged.

Group C — motion-initiated activation (cl. 5, 9, 21). Gogulski's travel/ready light scheme teaches sensing cart movement vs. standstill; Gretag's motion detector 74 (scale motion) is a motion sensor. Powering up on cart movement to conserve a rechargeable battery is a predictable power-management optimization — a classic "recognized design incentive" under KSR.

Group D — bags/containers in the receptacle (cl. 6, 11, 26). Gogulski expressly discloses the removable bag in the cart to eliminate separate bagging. The bag-height limitation of cl. 11 (top not above the lower mirror edge) is an unclaimed-for-reason design constraint that follows directly from using the mirror area as the detection reference — routine dimensioning.

Group E — optical disk image storage (cl. 10). Video storage on magnetic tape/disc or optical media was conventional CCTV practice as of 1988 (the surveillance references). The '819 specification itself treats removing a video storage disk from the module and plugging it into a reader as an obvious alternative (FIG. 17 discussion), which is strong evidence that the storage-medium choice was within routine engineering.

Group F — display of item description and price (cl. 12). Gogulski displays 38/40/42/44/50; Clyne aggregate display; Checkrobot CRT prompts.

Group G — timed recording, ~1/5 second (cl. 13, 14). Checkrobot claims 4–5, 8 (a predetermined time period established after a code-reader signal with rejection on timeout); Gretag's motion-detector-triggered comparison. Choosing a sampling interval and a short record duration to conserve storage is optimization of a result-effective variable with no unexpected result shown.

Group H — error indication (cl. 15, 16, 17, 27). The strongest part of the rejection.

• Cl. 27 (error record if placement is not accompanied by scanning) is directly and expressly disclosed by Ehrat cl. 1: the equipment "indicat[es] … if an article has been placed in or removed from said container but the data has not been read correctly."
• Cl. 15 (camera detects passage across the upper colored border) — substitution of Gogulski's/Checkrobot's light-plane breach detection with camera-based breach detection.
• Cl. 17 (sensors around the upper periphery) is literally Gogulski's "light-sensing system located on the cart's top perimeter" and Checkrobot's optical curtain at the security-zone inlet.

Group I — variable-speed playback (cl. 22). Slow-motion/variable-speed review is an inherent, standard capability of the video-record/playback art (US 4,237,483 / US 4,630,110 family of surveillance systems) and requires no more than use of a known device capability for its known purpose.

---

6. §102(e)/§103(c) trap on US 4,779,706 (flag)

US 4,779,706 is NCR's own "Self-service system," filed 1986‑12‑17 and granted 1988‑10‑25 (within one year of the '819 filing, so not §102(b)). It could only be §102(e) art via its pre-filing US application date. Because pre‑AIA §103(c) disqualifies §102(e)/(f)/(g) subject matter that was commonly owned by the same person (or subject to an obligation of assignment) at the time the '819 invention was made, and both are NCR applications by different inventive entities, this reference is probably unusable in a §103 combination absent proof of non-common ownership. Do not build the primary rejection on it. The same caution applies to any NCR-internal document; note also that the '819 specification names the NCR 7052 POS terminal, confirming the common corporate context.

---

7. Motivation to combine (the crux)

Four independent, record-grounded motivations:

1. Identical problem, articulated in the art and in the patent. Gretag, Gogulski and Checkrobot all expressly address unauthorized/unrecorded placement of goods into the customer's cart. '819 recites the same problem as an object ("safeguards against incorrect or incomplete recording of articles purchased"). Where the references and the patent attack the same problem, motivation is inherent.

2. "Known technique to improve a similar device in the same way" (KSR). Video surveillance to create a reviewable record of retail transactions was a known technique (US 4,237,483, US 4,630,110). Applying a known recording technique to a known self-scanning cart system to replace or supplement its weight/light-plane verification is the paradigm KSR case.

3. A finite set of predictable alternatives, with express trade-offs. The art names weight comparison (Gretag, Clyne, Checkrobot), light-plane/optical-curtain breach detection (Gogulski, Checkrobot), and human review of a printed challenge list (Clyne: print-out "used to challenge a dishonest customer") — all serving the same verification function. Weight-based verification has acknowledged failure modes (same-weight substitutions, bulk/loose items, tamper with a cart scale). Video is a predictable alternative with a reasonable expectation of success. KSR: "obvious to try" where the prior art identifies a finite number of identified, predictable solutions.

4. Express design incentive / market force — and a motivation for the detachable module specifically. Checkrobot's own specification criticizes Gretag because each of 50–100 carts must carry the full electronics and the carts cannot nest — a "major storage problem." That criticism, combined with Gogulski's rechargeable-battery disclosure, supplies a direct motivation to (a) make the cart electronics detachable/shareable, (b) locate the bulky/expensive parts in a module at one cart end, and (c) rack-charge the modules. '819's own objects (minimize labor costs, reduce front-end checkout area, speed checking out and bagging) are market forces that KSR recognizes as legitimate motivation.

5. Reasonable expectation of success / level of predictability. Modular interfaces (Ehrat cl. 9 connector; Clyne's removed-then-mated unit), split-field optics (a single-sided mirror is a ubiquitous optical element and is the very structure '819 discloses at element 47), battery-powered mobile terminals, and printer/CRT/display output were all mature in 1988. Nothing in the claim requires a result that the art would have regarded as unattainable.

---

8. Anticipated patentee rebuttals, and how they fare

---

9. Bottom line — per-claim §103 exposure

Overall: US 4,929,819 is, on the record, a textbook combinatorial patent. Its independent claims recite a known cart-based self-scanning system (Ehrat, Gogulski, Clyne), a known security-verification function (same references, plus Checkrobot), and a known recording/playback technique (US 4,237,483, US 4,630,110) — combined with a detachable-module packaging choice that the art itself expressly motivates. The single most defensible feature is the specific dual-tone-bordered cart-interior mirror fiducial arrangement of dependent claims 2, 7, 8, 15 and 23–25, not anything in independent claims 1 or 18. Note also the procedural posture: the eight citations were before the examiner, but the four family cites — including the two surveillance systems and the NCR self-service reference — appear to have been raised only in the JP/EP counterparts and may not have been considered in US prosecution; a validity challenge would look there first (subject to the §103(c) caveat on US 4,779,706).

Confidence/caveats: My disclosure characterizations of US 3,836,755, US 4,071,740, US 4,373,133, US 4,676,343 and (partially) US 4,838,383 are grounded in retrieved text. My characterizations of US 3,949,194, US 4,237,483, US 4,369,361, US 3,749,191, GB 2,068,132 B, US 4,630,110 and US 4,779,706 rest on the face of the record (title/assignee/date) and background knowledge, not on retrieved full text — I did not verify them within this session's tool budget, and the rejection theories above must be re-validated against their actual disclosures before being asserted. I make no representation about litigation history or the current chain of title.