US 5729697

I'll search for the specific patent number and any litigation involving it.

Let me run a couple more targeted searches for litigation and docket records.

US 5,729,697 — Verified Summary

I searched for the exact number 5729697 and filtered out similarly-numbered hits. Findings below are drawn from the authoritative patent text you supplied, corroborated by Google Patents, FreePatentsOnline, Espacenet, and uspto.report.

Bibliographic data

Abstract (verbatim): "The present apparatus is a mobile shopping cart that can automatically keep track of objects selected and carried in the cart and provide a customer using the cart information like the total price of the items carried. The cart has a radio frequency cart base station with a cart antenna connected to a radio frequency cart transceiver. Various novel embodiments of the invention produce a radio frequency field within the shopping cart that is used to communicate with radio frequency tags on objects carried in the cart. The tags contain information about the respective carried object to which the tags are attached. A program function executed by the cart circuit and/or computer controls the cart transceiver to interrogate radio frequency tags within the radio frequency field, i.e. inside the cart to gain information about the carried objects. Tags on objects external to the cart are not interrogated."

Independent claims

The patent has 11 claims, and exactly one independent claim — claim 1. Claims 2–11 all depend, directly or indirectly, from claim 1.

Claim 1 (independent) — plain language: A mobile shopping cart that contains, on board, a radio-frequency base station made up of a cart antenna, an RF transceiver, and a cart computer with memory. The antenna creates an RF field within the cart (not the whole store). An "interrogator" module run by the cart computer drives the transceiver to periodically poll RF tags inside that field, where at least some of those tags are attached to objects sitting in the cart. The cart memory holds a data structure that the interrogator periodically updates, with one entry per object in the cart. That data structure stores an object price for each entry plus a running total of those prices. Separately, each RF tag has a "paid memory location," and the cart antenna writes paid information into that location via cart signals after the shopper pays the running total.

Dependent claims (2–11), briefly:

• 2–3: Running total shown on a cart display, or printed as a bill/receipt on a cart printer.
• 4–5: Shopper pays via a cart input device; claim 5 specifies a credit card reader (swipe).
• 6–7: Cart passes a security base station that interrogates the tags to check the paid memory location; claim 7 adds an alarm if any tag lacks paid information.
• 8–11: Cart display shows store information — claim 9 a menu of item categories, claim 10 a sale-item menu, claim 11 a store map.

Note on the claim text: claim 1 as published reads "controls the can transceiver" — a facial typographical error in the granted text for "cart transceiver." I flag it rather than silently correcting it, per the literal-interpretation rule.

CAFC / litigation status

Based on my search, I found no Federal Circuit appeal docket in 2026 (or any year) naming US5729697, and no district court or ITC litigation record for it. The only 2026 CAFC and 2026 litigation material surfaced by search concerned unrelated patents (e.g., Netlist v. Micron, Contour IP v. GoPro) with no connection to 5,729,697. I cannot confirm a complete negative from the search tools available to me — a Westlaw/Docket Navigator/PACER query would be needed to state that with certainty. What is confirmed from the patent record is that the patent expired for failure to pay maintenance fees / term end on 2015-04-24, so it is not in force today.

Identifier collision warning

There is an unrelated Japanese patent numbered 5729697 (J-GLOBAL) for a locking screw ("緩み止めネジ"), application 2014090133, filed 2014-04-24. Per your instruction not to auto-correct identifiers, I note this is a distinct national registration sharing the same numerals — it is not US 5,729,697 and should not be conflated with it. Likewise, US6032127A is the continuation of this application (same inventors/assignee, same priority date), not a separate invention.

I'll research litigation involving US patent 5729697. Let me run several targeted searches.

Let me search litigation-specific databases.

Let me run additional targeted searches.

Litigation Search: US Patent 5,729,697 ("Intelligent Shopping Cart")

Bottom line

I found no evidence of any litigation — past or pending — involving US Patent 5,729,697. Based on the searches I ran, there are no known district court cases, ITC investigations, PTAB proceedings, or Federal Circuit appeals asserting or challenging this specific patent. I am reporting this as a negative finding, not a certification that none exists (see limitations below).

Sources searched and what they returned

Discarded false positives (per your instruction not to auto-correct or conflate numbers)

• JP Patent No. 5729697 — a Japanese "loosening-prevention screw" (緩み止めネジ), applicant/patent owner 廣澤 勲. This is a different, unrelated patent sharing a similar number string. Discarded.
• The Google Patents, Espacenet, and CN100416601C documents that cite "US5729697A" are prior-art citations, not litigation records. Discarded.

Why this negative finding is plausible (context, not proof)

Per the patent record you supplied:

• The patent is Expired – Lifetime, with an anticipated expiration of April 24, 2015 (20-year term from the April 24, 1995 filing date). Its family members (US 6,032,127 B1; CA 2,168,919 A1; JP 3784451 B2) share the 1995 priority date.
• Ownership changed over time: IBM Corp. → UNOVA, Inc. (1998) → Intermec IP Corp. (1999). Intermec is a company that did historically enforce its RFID/IP portfolio, so enforcement activity at some point is conceivable but not established by anything I found.

Limitations — please read

1. Search-tool limitation: I could not confirm a page-by-page check of the Unified Patents litigation caselist filtered to this patent, nor a PACER/Docket Navigator query (those typically require subscription access). My searches were web-search based.
2. Coverage gaps: Older cases (1998–2010) and cases that were filed and voluntarily dismissed quickly are often poorly indexed on free web search. A "no hits" result does not conclusively rule out a short-lived or sealed suit.
3. This is not a legal conclusion or a formal clearance/FTO search. It is a documented negative finding from public web sources.

Recommendation

For a defensible negative-clearance statement, run a paid-database search on:

• Docket Navigator / Lex Machina / PACER for "5,729,697" as an asserted patent, and
• PTAB / Unified Patents for IPR/PGR/CBM filings, and
• ITC EDIS for Section 337 investigations,

and confirm with the patent's litigation history as reflected in the USPTO Patent Center file wrapper / assignment records.

If you have a specific case number or party name you were expecting to find, send it and I will run a targeted search to confirm or refute it.

I'll verify the structured ODP block against public web sources before reporting.

The top hit is a red herring — it's about U.S. Patent 10,425,697 (Genius Sports), not 5,729,697. Let me run additional targeted searches.

Let me run a few final targeted checks on PTAB databases and the continuation patent.

PTAB / AIA Trial Proceedings — US 5,729,697 ("Intelligent Shopping Cart")

Proceedings overview

Total AIA trial proceedings on file: 0. The structured USPTO ODP block states plainly: "no AIA trial proceedings for this patent as of the most recent ingest." My web-search sweep (PTAB petitions documents, PTAB statistics, Unified Patents portal pages, CAFC/PTAB citations) surfaced no IPR, PGR, or CBM naming US 5,729,697 as the challenged patent — so the breakdown across active / claims invalidated / claims sustained / settled / institution denied is 0 / 0 / 0 / 0 / 0. The defensive posture this gives a defendant is therefore not "hardened by surviving IPRs" and not "already invalidated" — it is untested at the Board, and the more controlling fact is that the patent expired 2015-04-24 (anticipated term end, per the record you supplied). There is no live PTAB trial to cite, and no FWD to quote.

---

Red herrings I discarded (per the no-auto-correct rule)

• U.S. Patent No. 10,425,697 (Genius Sports Ltd. v. patent owner; Petition filed 2024-08-22; Kirkland & Ellis). The petition text literally says "The '697 has not been subject to any prior IPR or PGR petitions." This is a different patent — 10,425,697 shares only the trailing digits "'697" with 5,729,697. It is not a proceeding on your patent and must not be cited as one.
• JP Patent No. 5729697 ("緩み止めネジ," loosening-prevention screw), a distinct Japanese registration flagged in the earlier sections. Not this patent.
• US 6,032,127 B1 — the continuation of this application (same priority date, same inventors). Any future proceeding search should check it as a separate claim set, but no proceeding on it surfaced either.
• Intermec v. Palm (D. Del.) and Alien Technology v. Intermec (D. Minn.) — these are district court cases involving Intermec's data-capture portfolio. Neither is a PTAB proceeding, and neither asserts 5,729,697. Not evidence of AIA activity.

Why zero PTAB activity is plausible here

• Timing mismatch: the AIA trial regime (IPR from 2012-09-16; CBM/PGR from 2012-09-16) largely post-dates the enforcement window of a 1995-filed patent. By the time IPRs were routine, this patent was years from expiry, and it finally expired on 2015-04-24 for failure to pay maintenance fees / term end. There was little practical window and even less incentive for a challenger to spend $500K+ invalidating a patent nobody was visibly asserting.
• No assertion anchor: the earlier sections found no district court or ITC litigation asserting 5,729,697. IPRs overwhelmingly follow infringement suits. No suit → no petition. The absence of PTAB activity is the expected corollary, not an anomaly.

---

Strategic summary

Claim status: all 11 claims are UNTESTED, not canceled. There is no FWD, no certificate of cancellation, no adverse judgment. Nothing in the public record narrows, cancels, or confirms any of claims 1–11 at the Board. If someone today hands you a demand letter citing 5,729,697, you cannot say "claim 1 is dead" — that is false here. What you can say is that the patent is expired (term ended 2015-04-24), which undercuts any forward-looking injunction theory and limits damages to the pre-expiry period, and that the asserted claims have never been substantively tested in any adjudicative forum so far as the public record shows.

Estoppel landscape — effectively moot, but not because of AIA. § 315(e)(2) estoppel only arises from a petition that was instituted and reached an FWD. No FWD exists, so no statutory estoppel attaches to anyone on this patent. Practically, because there is no co-pending PTAB trial, a defendant retains the full universe of invalidity grounds — § 102/§ 103 art, § 112 positions, and § 101 subject-matter challenges, subject only to ordinary district-court/ITC scheduling and IPR time-bar rules (§ 315(b) one-year bar) if you decide to file your own petition. Note the usual catch: an IPR petition can only raise § 102/§ 103 on patents/printed publications, so § 112 and § 101 defenses stay in court regardless.

Pattern signals — none to read. No repeat petitioner, no defensive aggregator (Unified Patents et al.) in the chain, no PTAB appeals by the patent owner (the ownership lineage IBM → UNOVA → Intermec IP Corp. is an assignment chain, not an enforcement pattern). One data point is worth flagging: the '697 family has an extensive forward-citation trail (222 "Cited By" entries, including later shopping-cart/RFID patents), which is a signal of technological importance as a prior-art reference — not of litigation or PTAB heat on the patent itself.

---

Recommended next steps

1. If you're a defendant: state the position accurately — "No PTAB proceeding has ever been instituted or decided on US 5,729,697; no claims have been canceled; the patent expired 2015-04-24." Do not assert IPR estoppel, because there is no FWD to anchor § 315(e)(2).
2. If active litigation exists that I could not see: check the docket for an IPR/civil-action-stay motion, then confirm against PTAB E2E and the ODP API. As of this analysis, the ODP block and web sweep agree: nothing on file.
3. If you're considering filing your own IPR: it is a practical dead end — the patent is expired, so there is no live right to cancel and no controversy to resolve; the Board generally will not institute where the patent has expired and all claims have lapsed (there must be a live "claim to be canceled"). A reexamination or a court invalidity holding is the wrong tool for a lapsed patent, too. Redirect the effort to the continuation US 6,032,127 B1 in case any rights there remain live.
4. Verify via primary sources (the transparent version of the negative): PTAB E2E (https://ptacts.uspto.gov/ptabweb/), the USPTO Open Data Portal AIA API, and CourtListener's RECAP docket search for "5,729,697." My searches were web-based and could not page-check every PTAB record; the ODP block is the canonical list and it returns none.

---

Honest caveat: I found no PTAB activity, and I am reporting that as a documented negative — not a certification. I could not run a subscription PACER/Docket Navigator/PTAB-E2E query within this session, and short-lived, terminated, or unindexed proceedings are the classic blind spot for free web search. No proceeding numbers, panels, grounds, or FWDs are invented here, because none exist to report.

I'll research the assignment chain for this patent, starting with the recorded reassignments in the patent record.

Good lead on the Intermec/UNOVA relationship. Let me pin down the actual reel/frame numbers for this patent's assignments.

I hit the search-step ceiling before I could pull the reel/frame and correspondent fields from Assignment Center. Below I report exactly what the available record supports and flag the gaps rather than fill them.

---

Inventors

Pattern check — inventor departures: No evidence surfaced that either inventor left IBM within 12 months of the 1995-04-24 filing. Both are career IBM Research personnel and the invention reads as a research-conceived concept (RF-tagged cart), not a startup spinoff. Departure-within-12-months signal: not present (as far as determinable). I could not independently verify either inventor's employment end-date from primary sources, so I mark this not present/unclear rather than affirmative.

---

Original assignee

• Entity on the issued patent: International Business Machines Corporation (IBM), Armonk, NY.
• Primary business: Computers, software, and services. The invention arose inside IBM Research (Almaden + Watson), which is consistent with the two inventors' affiliations.
• Product embodying the claims? No evidence that IBM ever shipped an "intelligent shopping cart." This reads as a research/disclosure asset, not a product-backed patent — which is consistent with IBM divesting it ~9 months after issuance.
• Current status: IBM remains an operating company. IBM itself did not retain ownership; per the patent record it assigned the patent away on 1998-12-21, and the current assignee listed by Google Patents is Intermec IP Corp.

---

Assignment timeline

⚠️ Important sourcing caveat: The authoritative patent text you supplied (Google Patents legal-events feed) gives the dates, conveyance type, assignor, and assignee for each reassignment but does not include reel/frame numbers or the correspondent of record. My web verification was cut off before I could retrieve those fields from Assignment Center. I therefore leave reel/frame and correspondent blank rather than fabricate them. The three events below are grounded in the patent record; the reel/frame values are not obtained.

1995-04-24 (executed) / recorded 1995-04-24 — Reel/Frame: not obtained

• Conveyance: Assignment of Assignors' Interest (inventors → employer)
• Assignor: Mario Schkolnick; Robert J. von Gutfeld
• Assignee: International Business Machines Corporation
• Correspondent: not obtained
• Context: Standard pre-issuance invention assignment to the employer of record.

1998-12-21 (executed) / recorded 1998-12-21 — Reel/Frame: not obtained

• Conveyance: Assignment of Assignors' Interest
• Assignor: International Business Machines Corp.
• Assignee: UNOVA, Inc.
• Correspondent: not obtained
• Context: Portfolio divestiture/sale — IBM moved this asset (and, per contemporaneous Intermec history, a broader block of RFID patents) to an industrial-automation operating company, roughly nine months after issue.

1999-02-08 (executed) / recorded 1999-02-08 — Reel/Frame: not obtained

• Conveyance: Assignment of Assignors' Interest
• Assignor: UNOVA, Inc., a Delaware corporation
• Assignee: Intermec IP Corp., a Delaware corporation
• Correspondent: not obtained
• Context: Intra-corporate consolidation — UNOVA pushed the just-acquired RFID IP down into its Intermec IP holding subsidiary (Intermec being the RFID/barcode operating arm).

Not an assignment but relevant: 2015-04-24 anticipated expiration. Status = Expired – Lifetime. Also, the 1997-10-21 continuation (US08/955,467 → US6032127A) is a same-family, same-inventors filing, not a third-party transfer.

No records for this patent? No — the patent does have recorded post-issuance assignments (per the patent record). It does not have a later record terminating the chain at a defensive aggregator, and no further 2000s-era assignments appear in the Google Patents legal-events feed.

---

Timeline diagram

timeline
    title Ownership of US 5729697
    1995 : Filed Apr 24 by IBM
         : Inventor assignment recorded
    1998 : Issued Mar 17
         : IBM assigns to UNOVA Inc Dec 21
    1999 : UNOVA assigns to Intermec IP Corp Feb 8
    2015 : Patent expires Apr 24

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The chain moves from an operating giant (IBM) to an operating industrial company (UNOVA) to its subsidiary Intermec IP Corp. The "IP" name suffix is a naming tell only; there is no evidence of a no-product, registered-agent-address, single-purpose shell. Intermec is a genuine RFID/barcode manufacturer, and Intermec IP Corp. is its IP-holding arm, not an anonymous assertion vehicle. No reel/frame evidence retrieved, so this is a "not present on the naming test alone" call.

2. Known asserter in the chain — not present. Neither UNOVA, Inc. nor Intermec IP Corp. appears on the listed NPE lists (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation, Spangenberg entities). Intermec has been an operating-company plaintiff in RFID disputes historically, which is categorically different from the NPE pattern. Note: this is a general enforcement posture; I could not confirm any suit naming this patent (consistent with the prior litigation section in this file).

3. Repeat correspondent across the chain — unclear (data not obtained). This is the signal most damaged by my sourcing gap: the correspondent of record for the 1995, 1998, and 1999 recordings was not retrievable in the steps available. Without the correspondent names I can neither confirm nor rebut recurrence. Flagging as a gap to close.

4. Cascading transfers — not present (weak form only). There are two transfers in ~7 weeks (1998-12-21 and 1999-02-08), but they run IBM → UNOVA → UNOVA's own subsidiary, i.e., a sale followed immediately by an intra-corporate downstream assignment — not a chain of unrelated LLCs. No shared opaque correspondent address or common-principal LLCs are evidenced.

5. Pre-litigation transfer — not present. No infringement suit naming US 5,729,697 was found (see the litigation section of this analysis), so there is no 6-month-before-suit transfer to anchor this signal. The 1998/1999 transfers predate any conceivable assertion window by years.

6. Bankruptcy fire-sale — not present. IBM was not in bankruptcy; the 1998 transfer was a voluntary divestiture of non-core RFID IP, and the transferee (UNOVA/Intermec) was a going concern.

7. Privateering — not present / unclear. IBM transferred to a direct RFID competitor/operating company (Intermec), not to an entity asserting on IBM's behalf. No SEC-filing or EFF/Patent Progress evidence of a back-end privateering arrangement surfaced.

8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at an operating company.

Contextual note (must not be conflated): A search surfaced the INPADOC legal status of a different patent, US 6,307,349 B1, showing Intermec IP Corp. (Delaware) → UNOVA, Inc. (California), effective 2001-02-06, Reel/Frame 011485/0693, and a later merger of UNOVA into Intermec, Inc. (Washington), Reel/Frame 026000/0535. Those reel/frames belong to US 6,307,349, not to US 5,729,697, and I expressly decline to attribute them to this patent. They are cited only to illustrate the depth and complexity of the UNOVA↔Intermec ownership relationship in the 1999–2011 window.

---

Verdict

Operating-company assertion (with a stated caveat).

Justification: The full chain is reconstructible from the patent record — inventors → IBM (1995) → UNOVA, Inc. (1998-12-21) → Intermec IP Corp. (1999-02-08) — and every link is a going-concern operating company or its IP subsidiary, with zero matches to asserter directories and no shell-LLC cascade. That is the opposite of a troll chain; the current assignee (Intermec IP Corp.) is the IP arm of an RFID/barcode manufacturer that ships products in this space. Caveat: the "assertion" prong (suing actual competitors) is not confirmed for this specific patent — no suit naming US 5,729,697 was found. If a confirmed suit is required to use this verdict, the answer degrades to Insufficient data on assertion, though the ownership itself is not NPE-patterned. Verify at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (patent-number search) and https://assignment.uspto.gov/patent/index.html — pull the reel/frame and correspondent for the three recordings above; the correspondent for the 1998/1999 links is the single highest-value missing datapoint for confirming the absence of an NPE pattern.

I'll research the prior-art citations on the face of US 5,729,697. Let me run several targeted searches against patent databases.

I have the patent's own background discussion. Now let me pull the front-page "References Cited" (prior-art) list that the examiner/applicant cited on US 5,729,697.

Let me pull the front-page reference list from FreePatentsOnline and the patent PDF.

Scope and method note (important framing)

"Every patent citation for 5,729,697" can mean two different, non-overlapping sets — and only one of them is prior art:

1. Backward citations (prior art relied on / referenced by the '697 patent itself). These are the references the inventors incorporated and discussed in the Background, plus the examiner-cited "(56) References Cited" front-page list. These are the prior art you asked about.
2. Forward citations — the "Cited By (222)" list in the record you supplied. Those are later patents that cite 5,729,697 (e.g., US 5,933,813; US 6,123,259; US 6,484,939; US 2006/0289637 etc.). They issued after 1998 and therefore cannot anticipate or predate the '697 patent. They are evidence of the patent's downstream influence, not prior art. I exclude them from the § 102 analysis below.

Honest limitation up front: the full-text record you supplied and the search results I retrieved did not contain the front-page (56) "References Cited" table (Google Patents renders it as a table that the fetch dropped). I therefore cannot certify the complete examiner-cited list, and I will not invent entries to fill it. What I can state with confidence is the set of references expressly cited and discussed inside the '697 specification, which I set out below. If you need the certified (56) list, pull it from the granted-PDF front page or the USPTO Patent Center file wrapper.

---

Prior art expressly cited in the '697 specification (verified from the patent text)

---

§ 102 potential-anticipation analysis, by reference

A critical threshold point: for a reference to anticipate under § 102 it must disclose every element of the claim, arranged as claimed. Claim 1 is the only independent claim; its elements are (i) a cart-mounted RF base station (antenna + transceiver + computer/memory); (ii) an in-cart RF field (tags external to the cart are not read); (iii) an interrogator that periodically polls tags inside that field; (iv) a periodically updated data structure with a price per entry plus a running total; and (v) a "paid" memory location on each tag written by the cart after the shopper pays.

---

Claim-by-claim § 102 bottom line

• Claim 1 (independent): No cited reference anticipates it. Each of A–F is missing at least the combination of (a) the cart-local RF field that excludes external tags, (b) periodic re-polling that refreshes the tally, (c) the running-total data structure, and (d) the on-tag "paid" bit written by the cart after payment. The references collectively attack the individual elements (tags, carts, checkout, multi-tag ID), which is a § 103 obviousness posture — and indeed the '697 applicants distinguished this art in the Background by pointing to the absence of "constant periodic re-polling."
• Claims 2–3 (display / printer): not anticipated; Anders/Chenoweth show price totaling but not on-cart display/printing driven by periodic polling.
• Claims 4–5 (payment device / credit-card reader on cart): Chenoweth (C) is the nearest art (cart-attached checkout), but it is manual, so not anticipatory.
• Claims 6–7 (security base station checking "paid" tag bit; alarm): no cited reference discloses an on-tag paid bit interrogated at an exit gate — this was the applicants' anti-theft contribution. Not anticipated.
• Claims 8–11 (store info: category menu, sale menu, store map): none of the cited references disclose on-cart promotional/map display. Not anticipated.

---

Caveats you should carry forward

1. Source gap: I could not retrieve the certified (56) examiner-cited front-page list within this session; the references above are those I can verify from the '697 specification text and incorporated documents. Publication/issue dates marked "verify" (A, C, D) come from the patent's narrative, not a page-checked bibliographic record — confirm them against the granted PDF before relying on them for an invalidity chart.
2. Forward ≠ prior art: the 222 "Cited By" entries (including US 5,933,813, US 6,123,259, US 6,484,939, US 6,219,222-family items, and the Media Cart Holdings 2006/2007 publications) are post-1998 and legally cannot be § 102 art for the '697 patent. Don't let them be conflated with the references above.
3. Identifier discipline: per your rule, I have not auto-corrected any number. Note that US 6,032,127 is the continuation of this application (same priority date, same inventors) — it is family, not prior art; and the Japanese JP 5729697 ("loosening-prevention screw") is an unrelated registration sharing the numeral string.
4. Mootness: the '697 patent expired 2015-04-24, so any § 102 analysis is now retrospective (relevant to pre-2015 damages, licenses, or defensive history), not to live enforcement.

If you can supply the certified (56) front-page "References Cited" table (or the granted-PDF front page), I will map each examiner-cited reference to the specific claims with the same rigor.

I'll ground this analysis in the prior-art references the patent itself identifies, then verify them against public sources.

Let me verify the Chan application, which the patent expressly incorporates by reference.

Obviousness Analysis — US 5,729,697 ("Intelligent Shopping Cart") under 35 U.S.C. § 103

Analyst work product. This is a technical obviousness assessment based on the prior art identified in the patent's own specification plus public records I could retrieve. It is not a legal opinion, an invalidity contention, or a formal FTO/validity search.

---

0. What I'm working from (and what I verified)

The '697 specification itself identifies the relevant prior art by name. I verified each against public sources; all predate the 1995-04-24 priority/filing date, so all qualify as prior art under pre-AIA § 102(a)/(b)/(e):

Two clarifications that matter for the analysis:

1. Chan is a § 102(e) reference and incorporated by reference into the '697. Because the '697 spec expressly incorporates Chan ("This application is herein incorporated by reference in its entirety"), the multiple-tag interrogation protocol is effectively part of the '697's own disclosure. That cuts both ways: it does not defeat novelty by itself (it's the applicant's own incorporated material), but it also means the '697 cannot point to the interrogation protocol as the inventive contribution — the contribution, if any, must lie in the combination of that protocol with a cart-confined field, a periodically-updated running-total data structure, and a tag-based "paid" flag.
2. I did not see the '697's own "References Cited" list in the supplied text (the Google Patents capture I was given shows the "Cited By" list but not the examiner-cited prior art). So I cannot state from this record which of these five references were actually before the examiner. That is a genuine gap (see § VII).

---

I. Framework

Under Graham v. John Deere the analysis is: (a) scope and content of the prior art; (b) differences between the prior art and the claims; (c) level of ordinary skill; (d) secondary considerations. Under KSR Int'l v. Teleflex, a combination is obvious where the references are from the same field or a analogous field, the combination is of known elements according to known methods to yield predictable results, and there is an articulated reason (design incentive, market pressure, known problem) to make the substitution — without the need for the reference to explicitly teach the combination.

The claims here are structural/functional apparatus claims; claim 1 is the sole independent claim and every other claim (2–11) adds a discrete, well-trodden peripheral feature. That structural posture makes a § 103 case comparatively straightforward: the question is not whether any one reference teaches the whole cart, but whether the combination of a cart-mounted RFID reader, a periodically-refreshed tally, and a write-to-tag-payment-flag was predictable to a skilled artisan in April 1995.

---

II. Level of ordinary skill (POSITA)

A person of ordinary skill in this art as of the 1995 priority date would have: a bachelor's degree in electrical engineering, computer engineering, or computer science (or equivalent experience), plus roughly 2–4 years of experience in RF identification/tagging systems and/or retail POS/checkout automation, with working familiarity with (i) passive backscatter transponder systems (Baldwin), (ii) anti-collision/multi-tag interrogation protocols (Marsh, Chan), and (iii) supermarket checkout/POS architectures and EAS systems (Anders, Chenoweth). This is a mature, combination-oriented art — the specification itself concedes that "RF technology for recognizing items by means of RF tags has been well known for several decades."

---

III. Claim 1 — element-by-element mapping to the prior art

Claim 1 (with the granted text's "can transceiver" typo for "cart transceiver" flagged in the earlier section) breaks into six limitations:

(a) "A mobile shopping cart" — Chenoweth: the checkout device is "transported by a customer, preferably on a shopping cart." Marsh: cart containing transponders. Fully met.

(b) RF cart base station (cart antenna + cart transceiver controlled by a cart computer with cart memory) — This is the Baldwin architecture ("base station / tag communication device") applied on a cart. Baldwin discloses the base-station-plus-transponder pairing; Marsh discloses an interrogator (transmitter, receiver, processor/microprocessor) and its host computer; Chenoweth discloses a self-contained controller (12) with memory transported on the cart. Combining a processor+memory unit (Chenoweth) with an RF transceiver/front-end (Baldwin/Marsh) yields (b). The '697 spec expressly states such tags and base stations are "well known in the art" and cites Baldwin.

(c) "cart antenna capable of creating a radio frequency field within the shopping cart" — Marsh discloses the interrogator's field encompassing the cart's transponders. Chenoweth's device is on the cart. Confining the field to the cart interior is a range/power sizing decision, and the '697 itself frames it as the solution to Anders' acknowledged problem:

Anders: interrogation "of the objects on the store shelf appears to be redundant with a final interrogation done at a check out counter."

So the reference that supplies the cart-mounted field also supplies the reason to confine it — avoid reading shelf/other-cart tags (the "duplicate checking" objection).

(d) "interrogator… controls the cart transceiver to periodically interrogate RF tags… attached to an object in the shopping cart" — Anders discloses store antennas that "send out interrogation signals over various time intervals" (i.e., periodic interrogation); Marsh discloses repeated interrogation of cart transponders; Chan discloses the multi-tag protocol that makes reading a full cart feasible ("on the order of a second," per Chan and the CSIR/BTG announcement the '697 cites). The claimed "interrogator" is a programmed controller driving a transceiver — routine.

(e) "data structure in the cart memory that is periodically updated… entry for one or more objects… includes an object price for each of the entries and a running total of the prices" — The running total is the feature the '697's own background says the prior art lacked: "The prior art does not disclose or recognize any automatic way of indicating to the shopper how the total bill is affected as objects are added or removed from the cart." But note the specification simultaneously concedes that manual running-total mechanisms existed on carts ("Some prior art provides these mechanisms on the cart, however, this requires customer action"). Chenoweth discloses price/description lookup from a central computer and customer-visible display/receipt. The remaining gap — automatic, periodic, tag-driven tally — is closed by Anders + Chan, and the running-total display itself is old.

(f) "a paid memory location… on each of the radio frequency tags and paid information is written… after the shopper pays the running total" — This is the strongest candidate for a genuine point of novelty, and it maps onto Chenoweth's marking/unmarking device (20): Chenoweth deactivates a security device on each item when scanned and reactivates it if the item is put back. The '697 substitutes a written bit in tag memory for the magnetic deactivation. Chan expressly contemplates writing to tags ("commands exist which output information to the tags… held temporarily… remain until over written… or remain permanently"), and the '697 itself calls the tag/base-station art "well known." So while the specific "paid flag written by a cart antenna" is a refinement, it is the kind of substitution of a known mechanism (tag data field vs. magnetic EAS state) for a known purpose (post-payment lockout), yielding a predictable result.

Net: every element of claim 1 is disclosed or rendered obvious by Chenoweth (cart-borne, pay-at-cart, security-at-exit, display/receipt) taken with Anders/Marsh (RF tagging + periodic interrogation in a cart) and Chan (multi-tag read + write-to-tag). The only element with real daylight is the combination of "periodically re-interrogated running total" + "paid flag in tag memory" — and the specification's own admissions do most of the work against non-obviousness there.

---

IV. Specific combinations and the motivation to combine

Combination 1 (primary): Chenoweth + Anders + Chan

The closest-art backbone is Chenoweth — it already places a wireless, memory-bearing, card-accepting, receipt-printing checkout device on a shopping cart and already has an exit security sensor. Its admitted shortcoming is that "each item selected has to be manually logged in by the shopper" and returned items must be manually logged out. The '697's stated objects — automatic tracking, running total, eliminating checkout facilities — are exactly Chenoweth's problem set.

Anders supplies the substitution of an automatic RF interrogation field for Chenoweth's manual scan, and additionally supplies periodic interrogation. Chan supplies the enabling protocol for reading many tags in one cart field quickly.

Motivation (KSR-aligned):

• Same field: both are supermarket checkout automation.
• Known problem in the references themselves: Chenoweth's manual scanning is slow/error-prone; Anders flags the redundancy of shelf + checkout re-interrogation. Solving the former (automate) and avoiding the latter (confine the field to the cart) are the two moves the '697 makes.
• Predictable result: replacing a manual scan with an RF tag read is the canonical automation substitution; the CSIR/BTG public demonstration (cited in the '697 background: ~50 items polled/registered in 1 second) shows the art had already recognized RF could replace bar-code scanning at checkout.
• "Preferably on a shopping cart" in Chenoweth is a pointer, not just a suggestion — it places the reader on the cart and thereby makes the cart the natural locus of the field.

Combination 2: Marsh + Anders + Chenoweth

Marsh supplies the cart-with-interrogator concept and the multi-transponder anti-collision technique; Anders supplies the tag/pricing/inventory layer; Chenoweth supplies the pay-at-cart, display, and exit-security layer. This is essentially the same combination with Marsh, rather than Chan, providing the multi-tag read. The '697 criticizes Marsh (repeated re-transmission degrades as cart fill increases; interrogator is at the "till"), which is itself an admission of what Marsh teaches and a design incentive to move/interrogate differently — a classic "known technique improved" posture under KSR.

Combination 3: Anders + Chenoweth only (for claims 2–11)

Even without Marsh/Chan, Anders (RF tags, periodic store interrogation, cart/basket tag) + Chenoweth (cart-borne device, display, card reader, receipt printer, exit sensor) renders most of the dependent claims obvious, because those claims are peripheral consumer features Chenoweth largely discloses.

---

V. Dependent claims (2–11) — all fail on the same references

Claims 2–5, 8–11 are, in substance, Chenoweth's own feature list re-expressed with an RF front end. Under KSR they add nothing that changes the § 103 outcome.

---

VI. Secondary considerations and how they'd be answered

• Commercial success / long-felt need: The background asserts a long-felt need to eliminate checkout lines, and the later industry (e.g., the forward-citation trail of 222 "Cited By" entries noted in the earlier sections) shows the market moved this way. But to rebut obviousness, any commercial success or licensing must nexus to the claimed features; the '697's own spec attributes most elements to known art, weakening nexus for the broadest claims. Note also that the patent's forward-citation trail is evidence of technological importance as a reference, not of non-obviousness of the claims.
• Teaching away: The '697 argues Marsh's approach degrades as cart fill increases and that Anders requires duplicate checking. Those are criticisms of specific implementations, not a teaching away from the combination; indeed they are the incentive to arrive at the '697's design. Chenoweth's manual-scan architecture does not teach away from RF; it merely predates the RF substitution.
• Unexpected results: None are disclosed that are tied to the claims. The stated benefits (speed, no re-handling, fewer errors) are the expected consequences of automating the checkout, several of which Chenoweth already claims.
• Prosecution/double-checking caveat: Because Chan was incorporated by reference, the applicant effectively conceded the multi-tag interrogation protocol as non-novel; whatever survived allowance presumably rested on the cart-confined field + periodically-refreshed tally + write-to-tag payment flag. That narrow basis is where a § 103 contest would focus, and where the references above are most directly on point.

---

VII. Honest limitations and caveats

1. I could not retrieve the '697's examiner-cited "References Cited" list from the supplied Google Patents capture (it shows "Cited By," not the front-page art). I therefore cannot say which of these five references were actually before the examiner, and thus cannot assess whether any were "newly discovered"—a fact that would materially strengthen a § 103 challenge. That requires the USPTO file wrapper (Patent Center) or the printed front page.
2. I do not have the full claims of EP 0 494 114 or EP 0 623 904 — I worked from their abstracts, published descriptions, and (for Marsh) the Norwegian family member NO 310262 B1. The claim-level mapping to Chenoweth's device 11 elements (controller 12, card reader 14, receipt generator 16, scanner 18, marking/unmarking device 20, keypad 22, display 23) is well-supported; finer details should be confirmed against the EP specification.
3. Claim construction matters. Claim 1 uses functional language ("interrogator," "periodically," "radio frequency field within the shopping cart") that invites disputes over scope (e.g., whether a field that "slightly penetrates the walls" by 2–8 inches, as the '697's FIG. 4 embodiment teaches, is still "within" the cart). This analysis takes the claims at face value; a narrower construction could change the specific mapping but not the overall combination logic.
4. This is a § 103 analyst assessment, not an invalidity opinion. It identifies plausible combinations and motivations; it does not establish anticipation, and it does not weigh claim-construction or secondary-consideration evidence in the manner a court or the Board would.
5. Identifier discipline (per instructions): I did not auto-correct anything. The '697's claim 1 "can transceiver" is reported as a facial typo for "cart transceiver," not silently fixed. Chan's US 5,550,547 is identified by its verified prosecution lineage from Ser. No. 08/304,340 (filed 1994-09-12) as the incorporated protocol reference; I did not conflate it with any other '547-family number.

---

Bottom line. US 5,729,697's claim 1 is, on this analysis, vulnerable to a § 103 challenge as a predictable combination of Chenoweth (cart-borne, pay-at-cart checkout device with display, card reader, receipt printer, and exit security sensor) in view of Anders/Marsh (RF tags + periodic interrogation of cart contents) and Chan (multi-tag identification with tag-write capability). The strongest non-obviousness redoubt is the combination of the periodically re-interrogated running total with a "paid" flag written into tag memory — but the specification's own admissions (tags/base stations "well known"; manual running totals known; the multi-tag protocol incorporated from Chan) substantially erode even that. Independent claims 2–11 would fall either with claim 1 or on Chenoweth alone.