US 4922434

I'll search for this specific patent number as requested, focusing on authoritative sources.

Search Results for US Patent No. 4,922,434

Important note on the search scope: I searched for the literal identifier "4922434" as instructed. Results confirm the U.S. patent grant. I found no CAFC (Court of Appeals for the Federal Circuit) 2026 dockets referencing this patent — it appears in no active appellate litigation. I also note that the same numeric string appears in unrelated records (e.g., JP4922434B2, a Japanese data-write-control patent, and a Brazilian merchant-registry entry), which I have excluded as different identifiers/domains and not the US patent you asked about.

Bibliographic Data (US 4,922,434)

Sources: Google Patents, https://patents.google.com/patent/US4922434/en ; USPTO Report, https://uspto.report/patent/grant/4922434

Abstract (verbatim)

"Linear interpolation is applied to each pair of placement coordinates in a data driven surface mount component placement robot in order to eliminate component misplacement due to board artwork dimensional inaccuracies. The linear interpolation algorithm is derived by measuring the actual dimensions of the artwork of a given board by an integrated vision system via a camera attached to the robot arm."

Plain-Language Overview of the Invention

The patent addresses a classic surface-mount assembly problem: a robot places components at nominal CAD coordinates, but the printed wiring board's artwork has shrunk or stretched by a few mils during manufacture. Over a large board, that distortion causes components to be misplaced. Rather than mechanically re-spacing parts (as in the prior-art Caccoma "batch compensator," US 4,342,090), Fule's invention does the correction in software.

The robot's arm-mounted camera photographs artwork registration rings at their nominal locations, the controller measures the actual distance between them, and compares that to the nominal distance stored in the placement file. The ratio yields an interpolation factor, which is then applied to every nominal component coordinate before the robot picks and places each part.

There are two independent claims: claim 1 and claim 4.

Independent Claim 1 — Software-Only Interpolation System

A component placement system comprising:

• (a) a robot arm and (b) optical means (e.g., arm-mounted cameras) that selectively generate signals;
• (c) a board mounting station holding a board with component sites and at least two registration marks;
• (d) storage means holding three data types: nominal position of the registration marks, nominal distance between at least one pair of marks, and nominal position of the component sites;
• (e) optical control means that retrieves the mark position data, moves the robot arm so the optical means is positioned over the nominal location of successive marks, and triggers a signal at each;
• (f) interpolation determination means that computes the measured distance between the mark pair and compares it with the nominal distance to develop at least one interpolation factor; and
• (g) robot arm control means that, for each component site, applies the interpolation factor to that site's nominal data to produce corrected site data, then drives the arm to pick up and place the component at the corrected position.

Key point: correction is computed per-site from stored nominal data — no mechanical re-spacing hardware is required.

Independent Claim 4 — Vision-Error-Corrected Interpolation System

Claim 4 repeats the claim 1 architecture (robot arm, mounting station, storage means, optical means, interpolation determination means, robot arm control means) but adds a vision error correction factor into the storage means. Under element (e), the interpolation determination means must:

1. determine the measured distance between the registration marks;
2. adjust that measured distance using the vision error correction factor; and
3. compare the adjusted measured distance against the nominal distance to develop the interpolation factor.

Key point: this claim accounts for the fact (stated in the specification) that robot vision systems are repeatable but not perfectly accurate — their reading error is calibrated out before the interpolation factor is computed. The spec describes deriving this factor by comparing repeated robot-vision measurements against a Coordinate Measuring Machine or laser interferometer (stated accuracy ±0.0005 mil).

Dependent Claims 2 and 3 (for completeness)

• Claim 2: limits the board to three registration marks, with storage holding nominal positions for all three.
• Claim 3: the marks form two pairs, each pair defining a line perpendicular to the other (i.e., orthogonal X and Y baselines).

Cited Prior Art (as listed on the face of the patent)

US 4,342,090 (Caccoma, IBM, batch chip placement); US 4,628,464 (Westinghouse, robotic component mounting); US 4,641,257 (Canon, alignment measurement); US 4,720,635 (Disco Abrasive, automatic alignment); US 4,723,221 (Nippon Kogaku, semiconductor manufacture); US 4,737,845 (Hitachi, surface-mount loading); US 4,780,617 (Nippon Kogaku, successive chip alignment). The specification also distinguishes Watanabe US 4,681,452 and Peterson US 4,651,203 by name in its background discussion.

Uncertainty / Caveats

• CAFC dockets: I found no 2026 (or any) Federal Circuit docket activity for US 4,922,434. If you need a definitive litigation check, that should be confirmed directly against PACER/CM-ECF and the CAFC's public docket, which my search tools do not directly index.
• Assignee chain: Google Patents and uspto.report both list Control Data Corporation as assignee; no later reassignment to a successor entity (e.g., Ceridian) appears in the records I retrieved, though such corporate history may exist outside the patent record.
• Claim text: the overviews above are paraphrases grounded in the verbatim claim text reproduced in the full patent document provided; consult the official USPTO/Google Patents claim text for exact wording.
• Typographical quirks in the original (e.g., "registraton marks," "optica means," "X(vision error) = 15.205−15.204 = 0.0010") are artifacts of the source document and have not been auto-corrected.

I'll research litigation involving US Patent 4922434 across patent litigation databases.

Let me search more targeted litigation sources.

Let me check a few more specialized litigation databases directly.

Litigation Search: US Patent 4,922,434

Bottom line: I found no record of any litigation involving US Patent 4,922,434.

Summary of the search effort

I searched for the specific patent number across patent-litigation-oriented sources and general web indexes, using the exact identifier "4922434" and the formatted "4,922,434," paired with terms including "litigation," "infringement," "lawsuit," "complaint," "district court," and "Westlaw," as well as the patent's identifying details (title, inventor, assignee). The searches returned:

Notably, the "Control Data" hits that surfaced involved different patents — e.g., Technitrol, Inc. v. Control Data Corp., 550 F.2d 992 (4th Cir. 1977) (Sharpless-Eichert inventors; ENIAC-related), Scam Instrument Corp. v. Control Data Corp., 458 F.2d 885 (7th Cir. 1972), and Control Data Corp. v. IBM, 421 F.2d 323 (8th Cir.). None of these involves the '434 patent, and I have not attributed them to it.

Caveats on the negative finding

I want to be explicit about the limits of this conclusion rather than state it as certain:

1. No case docket, complaint, PTAB/CBM record, or CAFC appeal referencing US 4,922,434 appeared in the results. On the available evidence, this appears to be a patent that was never asserted in litigation.
2. This is consistent with the patent's profile. It is a 1990-issued patent (filed 1988-06-16, priority 1988-05-25) assigned to Control Data Corp., directed to linear interpolation for a component placement robot (the AdeptOne robot with Adept XGS vision system is named as suitable hardware). It expired for failure to pay maintenance fees (lapse event recorded 1998; final discontinuation entry 2018-01-30 per the legal-status record). A patent that lapsed in the late 1990s left only a ~7–8 year assertion window, which would be unusual for any substantial infringement campaign.
3. I could not directly query the sources you named. I was unable to load the Unified Patents litigation case list (portal.unifiedpatents.com/litigation/caselist), PACER, or the CAFC docket directly within this session. My conclusion rests on the searches that did execute. If you need a fully authoritative negative, the following should be checked directly:
   • Unified Patents Portal – Litigation Case List: https://portal.unifiedpatents.com/litigation/caselist (search patent number 4922434)
   • PACER (all district courts; search "4,922,434" as a patent-in-suit)
   • USPTO Patent Trial and Appeal Board (PTAB) and CAFC docket searches
   • Lex Machina / Docket Navigator (commercial litigation databases)

Requested table

Because no litigation is known, the fields you asked for would be:

If you need this verified as a certified negative, I can attempt additional targeted searches (e.g., Docket Navigator, Lex Machina, Darts-IP) — but I want to avoid overstating confidence: based on everything retrievable here, there is no known litigation involving US Patent 4,922,434.

I'll verify the structured data against public sources before concluding.

Proceedings overview

Zero AIA trial proceedings have ever been filed against US 4,922,434 — the USPTO Open Data Portal returns no IPR, PGR, or CBM for this patent as of the most recent ingest (0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials), and independent searching surfaced no unindexed filings either. The bottom line for a defendant is not about PTAB outcomes at all: this patent lapsed on 1998-05-06 for failure to pay the 7.5-year maintenance fee, roughly 14 years before the first IPR was ever filed, so all four claims (1–4) stand exactly as issued — never tested, never narrowed, and not enforceable for any infringement occurring after the lapse date.

Structured-data reconciliation

Procedural timing caveat (the real reason there is no PTAB record): AIA trials became available 2012-09-16. This patent's enforceable life ended 1998-05-06. The filing window and the patent's life never overlapped.

Key dates for US 4,922,434

• 1988-05-25 — Canadian priority date
• 1988-06-16 — US application 07/207,691 filed (Control Data Corp.; inventor Andras G. Fule)
• 1988-07-05 — assignment recorded (effective 1988-06-23)
• 1990-05-01 — patent granted; 4 claims issued
• 1993-07-12 — 3.5-year maintenance fee paid
• 1998-02-14 — maintenance fee reminder mailed
• 1998-05-03 — "Lapse for failure to pay maintenance fees"
• 1998-07-14 — "Lapsed due to failure to pay maintenance fee," effective 1998-05-06
• 2008-06-16 — anticipated expiration (20-year date) recorded, moot given the 1998 lapse
• 2018-01-30 — status notice: "PATENT EXPIRED DUE TO NONPAYMENT OF MAINTENANCE FEES UNDER 37 CFR 1.362"

Claim-by-claim status (no IPR ever touched these)

Because there was no AIA trial, no claim-level invalidation exists to cite, and no FWD exists to quote. Anyone representing otherwise (e.g., "claims 1–5 were canceled") is misrepresenting the record — note this patent has only four claims, so claims 1–5 could not have been canceled in any event.

Strategic summary

Which claims are canceled vs. sustained vs. untested. None are canceled; none are narrowed by any PTAB or court ruling. Claims 1–4 are administratively intact but legally dead as an enforcement tool. The expiration event of record is 1998-05-06 ("Lapsed due to failure to pay maintenance fee"). Under 35 U.S.C. § 271, infringement requires making, using, offering for sale, selling, or importing during the term. No accused conduct after 1998-05-06 infringes. Even if some pre-lapse conduct existed, 35 U.S.C. § 286 caps pre-filing damages at six years, and a suit filed today (2026) could recover nothing — any actionable window closed in the early 2000s at the latest. Laches and § 282 defenses add further insulation.

Estoppel landscape — § 315(e)(2) is a non-issue. With zero petitioners, no party is estopped from anything, and there is no IPR record to work from. Two practical consequences: (a) a defendant has nothing to inherit (no FWD holding claims unpatentable, no prior-art grounds the Board already blessed); and (b) a defendant also has nothing to avoid (no § 325(d) "same art" risk, no estoppel trap). The ordinary prior-art arsenal is untouched — most notably, the patent's own cited art (US 4,342,090 Caccoma, US 4,651,203 Peterson, US 4,681,452 Watanabe, US 4,738,245 Hitachi, US 4,628,464 Westinghouse) and any pre-1988-05-25 printed publications remain fully available in a district court § 102/§ 103 challenge, as does § 112 and the 1998 lapse itself.

Pattern signals — none exist. No repeat petitioners (no Unified Patents, no defensive aggregator, no serial-filer IPR campaigns), no Patent Owner appeals to the CAFC (there is nothing to appeal), no parallel litigation trail surfaced. This is unsurprising: by the time the AIA regime arrived, the patent had already been abandoned by its owner for over a decade. Its continued citation in later patents is a prior-art citation signal, not an assertion signal — the patent is cited as background art (compare the similar-documents cluster: US 4,980,971; US 5,086,556; US 4,999,764; US 5,216,804), not litigated.

Recommended next steps

1. If you received a demand letter citing US 4,922,434, do not build an IPR strategy. The patent expired 1998-05-06 for nonpayment of maintenance fees under 37 CFR 1.362. Verify the fee/lapse record yourself at USPTO Patent Center (application 07/207,691) and on the Google Patents record for US4922434A, which reproduces the legal-events table: "1998-07-14 — Lapsed due to failure to pay maintenance fee, effective date 1998-05-06."
2. Treat the demand as a Rule 11 / § 285 exposure point. A pre-suit demand demanding royalties on an expired, lapsed patent — with no good-faith basis to distinguish pre- and post-1998 conduct and no possible § 286 recovery — is the kind of letter that supports a declaratory-judgment action and an exceptional-case motion. Preserve the letter and any pre-suit communications.
3. If a case is nevertheless filed, the dispositive motions are (i) failure to state a claim / no actionable infringement post-1998-05-06 and (ii) § 286 time bar for any earlier conduct — not an IPR. Note also that the Canadian counterpart CA1314085C shows the same "Expired – Fee Related" status, so there is no foreign counterpart pressure either.
4. If your real interest is the technology, not this patent, the operative prior art is the Caccoma/Watanabe/Peterson cluster cited on the face of the patent, plus the patent's own extensive specification disclosure (registration rings integral to the artwork, vision-error calibration against a coordinate measuring machine, per-component interpolation per FIG. 3 block 49). That material is now decades-old public-domain prior art usable against other patents in this space.
5. No PTAB milestone tracking is warranted — there is no institution decision deadline, no oral hearing, and no statutory FWD due date to calendar, because there is no proceeding. For completeness, the authoritative negative can be re-verified at PTAB E2E / PTAB Center; if a "US 4,922,434" hit appears there in the future, treat it with skepticism until you confirm the patent number byte-for-byte — searching on "'492" returns US 7,774,492 hits and searching "'434" returns US 8,575,434 and US 8,359,501 hits, exactly the misidentification trap this record is prone to.

Confidence statement: High confidence that no AIA trial proceeding has ever been instituted on US 4,922,434 — the ODP structured block is empty, and targeted searching produced only same-number-suffix false positives from unrelated patents. I found no Federal Circuit or district court decision adjudicating this patent's claims. What I cannot rule out from the sources available to me is a non-AIA post-issuance event (e.g., a pre-1998 ex parte reexamination or a 1990s district court case) that is not reflected in the Google Patents "publications" list; if that matters to your matter, pull the full file wrapper from Patent Center rather than relying on the negative inference here.

I'll start by verifying the assignment record and cross-referencing the original assignee's corporate history.

Now let me check for litigation and any post-issuance transfers.

Let me check the assignment record directly and search for any trace of this specific patent in litigation or transfers.

Inventors

Nothing unusual to report on the inventor side: this is the ordinary single-employee patent. The invention was filed very near the end of CDC's life as a computer manufacturer, so no meaningful "inventor departure within 12 months" analysis is possible — by the time the patent issued (1990-05-01) CDC was already divesting. I found no evidence of Fule's subsequent employment, and I will not speculate.

Nothing in the file supports a "all inventors departed before a fire-sale" pattern — there is one inventor and no sale.

Original assignee

Control Data Corporation (CDC), Minneapolis/Bloomington, Minnesota. State of incorporation: Delaware (per the 8th Circuit caption in Ceridian Corp. v. SCSC Corp., 212 F.3d 398 (8th Cir. 2000), where Ceridian is styled "a Delaware corporation, as successor and assignee of Control Data Corporation").

• Primary line of business: mainframe/supercomputer manufacturing (CDC 1604, 6600, 7600, Cyber line), plus data storage (Imprimis) and information services; it also ran a Computer Integrated Manufacturing (CIM) division selling CAD/CAM and automated electronics-assembly integration. The patent's own specification describes the invention as implemented on an off-the-shelf AdeptOne robot with the Adept XGS Vision System of Adept Technology, San Jose, Calif. — i.e., the claimed subject matter was a software/data-driven placement method intended to run on a purchased commercial robot, not a CDC-built robot. I found no evidence of a specific CDC product shipped embodying claims 1–4, and no product literature or SEC disclosure linking this patent to a CDC product. Flagging that as unknown rather than denying it.
• Current status: dissolved/broken up. Effective 1992-06-01 CDC split into Ceridian (services; now Dayforce, Inc.) and the spun-off Control Data Systems, Inc. (computer operations, later acquired by Syntegra/BT). General Dynamics bought the former CDC defense-electronics unit Computing Devices International in December 1997. CDC as a corporate entity no longer exists.

Assignment timeline

One recorded assignment only. Everything else in the Google Patents legal-event table is maintenance-fee administration, not a conveyance.

• 1988-06-23 (executed) / recorded 1988-07-05 — Reel 004909/0007

  • Conveyance: Assignment ("ASSIGNMENT OF ASSIGNORS INTEREST")
  • Assignor: Fule, Andras G.
  • Assignee: Control Data Corporation, 8100 34th Ave. South, Minneapolis, MN (Minnesota)
  • Correspondent: Not determinable from the sources I could reach. Google Patents' legal-events table reproduces the reel/frame and the "ASSIGNMENT OF ASSIGNORS INTEREST; ASSIGNOR: FULE, ANDRAS G." free-text but omits the correspondent of record, and I could not query US-PTO Assignment Center directly to read the correspondent field off the recorded cover sheet. The nearest datapoint I could find — a different CDC assignment, reel 5884/96 (Hood et al. → Control Data Corp., executed 1991-10-15, recorded 1991-10-21, per Plainsite) — lists Merchant & Gould, P.C., 3100 Norwest Center, Minneapolis, MN 55402 as correspondent. That firm was clearly doing routine CDC employee-assignment recording work in this era, but I have not verified it is the correspondent on 004909/0007 and I am not treating it as a finding. No repeat-correspondent signal can be scored either way.
  • Context: Ordinary employee-inventor assignment to the original employer (the condition of employment), not an acquisition, fire-sale, reorg, securitization, or transfer to an asserter.

• No further assignments are recorded. After 1988-07-05 the legal-event record contains only fee events:

  • 1990-10-31 — FEPP (payor number assigned, large entity)
  • 1993-07-12 — FPAY, year-of-fee-payment 4
  • 1998-02-14 — REMI maintenance-fee reminder mailed
  • 1998-05-03 — LAPS (lapse for failure to pay maintenance fees)
  • 1998-07-14 — FP, lapsed due to failure to pay maintenance fee, effective 1998-05-06
  • 2018-01-30 — STCH, patent expired due to non-payment (37 CFR 1.362)

  So the patent died in the hands of its original assignee in 1998, roughly at the 8-year fee window, about six years after CDC's 1992 break-up. No NPE ever took it. The only family member is the Canadian counterpart CA 1314085C (priority 1988-05-25, granted 1993-03-02), also CDC-assigned.

Caveats on cross-referencing: several search hits for a "'434 patent" in litigation involve different patents — Robert Bosch LLC v. Pylon Mfg. Corp. (wiper-blade '434), Baxter/Fresenius (touchscreen '434), Altair Logix v. SparkFun (US 6,289,434), and Netlist's '434. None of these is US 4,922,434. I found no litigation, ITC action, or RPX/Unified Patents asserter listing naming US 4,922,434.

Timeline diagram

timeline
    title Ownership of US 4922434
    1988 : Priority application filed in Canada
         : US application filed 07 207 691
         : Assigned to Control Data Corp reel 004909 0007
    1990 : Patent issued
    1992 : Control Data split into Ceridian and CDS
    1998 : Lapsed for unpaid maintenance fee
    2018 : Expiration recorded by USPTO

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The only recorded conveyance (reel 004909/0007) runs from an individual inventor to the operating employer. There is no downstream transfer to any "IP / Patents / Licensing / Holdings / Ventures" entity, no Delaware/Texas single-member LLC, no registered-agent address in the chain.

2. Known asserter in the chain — NOT PRESENT. Assignee of record throughout is Control Data Corporation and, by corporate succession, Ceridian/Dayforce. Named assets are entirely absent from the Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, or Spangenberg families, and no RPX/Unified Patents high-frequency-plaintiff listing names this patent or its owner.

3. Repeat correspondent across the chain — UNCLEAR / NOT SCORABLE. There is only one link, so recurrence is impossible by construction; and the correspondent field for reel 004909/0007 could not be retrieved from the sources available to me. Note for completeness: the adjacent CDC assignment reel 5884/96 used Merchant & Gould, P.C. (Minneapolis) — a legitimate general-practice IP firm doing CDC's employee assignments — which is the opposite of a troll-side tell. No signal.

4. Cascading transfers — NOT PRESENT. Zero post-issuance assignments in 30+ years; no chained LLCs, no shared correspondent addresses, no sub-24-month sequences.

5. Pre-litigation transfer — NOT PRESENT. There is no infringement suit naming this patent at all, so no transfer could be "within 6 months before" one.

6. Bankruptcy fire-sale — NOT PRESENT (but adjacent restructuring context noted). CDC's 1988–1992 contraction (ETA shutdown, $680.4M FY1989 loss, sale of Imprimis, the 1992 two-way split) was a restructuring/divestiture, not a Chapter 7/11, and no assignment of this patent appears in the record as part of any asset sale. Consistently, CDC/Ceridian simply stopped paying the 8-year maintenance fee and let it lapse in 1998.

7. Privateering — NOT PRESENT. No operating company → NPE transfer, no assertion against competitors, no SEC/EFF/Patent Progress coverage.

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT, Unified, or OIN. The patent was abandoned by attrition (fee lapse), which is a different (and non-neutralizing) outcome — the art is simply dedicated to the public.

Verdict

Insufficient data — in the rubric's precise sense of "only the original assignment."

The entire chain is a single recorded conveyance: Andras G. Fule → Control Data Corporation, executed 1988-06-23, recorded 1988-07-05 at reel 004909/0007. No assignment, security agreement, merger, or name change was ever recorded thereafter; the only post-1988 events are fee payments and the 1998-05-06 lapse for non-payment (recorded 1998-07-14, code FP), with formal expiration logged 2018-01-30. All eight NPE signals are therefore not present or not scorable, and there is no evidence of assertion activity — so this is not an NPE chain, an operating-company assertion campaign, or a defensive-aggregation outcome. It is simply a dead 1988 CDC patent that lapsed in the original owner's hands.

Verify at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (search patent number 4922434) or the legacy interface https://assignment.uspto.gov/patent/index.html — I recommend reading the correspondent field on reel 004909/0007 directly there, since that is the one data point I could not confirm from the accessible sources. The full legal-event history is at https://patents.google.com/patent/US4922434/en#legalEvents.

I'll research this patent and its cited prior art. Let me run several searches in parallel.

Let me search for the remaining cited references and the two prior-art patents discussed in the specification body.

Prior-Art Analysis — US 4,922,434 (US4922434A)

Note on identifiers: I have interpreted every patent number literally and have not auto-corrected any. The authoritative text (provided) contains OCR artifacts (e.g., "registraton," "opticla," "osition"); I reproduce identifiers as given.

1. The patent at issue

Claim landscape (used for the §102 mapping):

• Claim 1 — System: (a) robot arm; (b) optical means on the robot arm; (c) board mounting station; (d) storage means holding (i) registration-mark nominal position data, (ii) nominal distance data, (iii) nominal component-site data; (e) optical control means that moves the arm to the nominal positions of consecutive marks; (f) interpolation determination means that measure a distance between marks and compare it to the nominal distance to develop an interpolation factor; (g) robot arm control means that, for each component site, applies the factor to that site's nominal data to generate corrected data and controls the robot to pick and place there.
• Claim 2 — three registration marks.
• Claim 3 — two perpendicular pairs of marks.
• Claim 4 — as claim 1, but the storage means additionally holds a vision error correction factor, which the interpolation determination means uses to adjust the measured distance before comparison.

2. Backward citations of record (7 references)

These are the references listed under "Patent Citations (7)" on the Google Patents record, marked per that record's legend as "* Cited by examiner."

2.1 US4342090A — "Batch chip placement system"

• Citation: US 4,342,090 A; issued 1982-07-27; filed 1980-06-27; International Business Machines Corp.; inventors Caccoma, O'Neill, Tappen, Koestner. Corresponding EP 0042977 A1.
• Description: A batch chip placement system that senses the X and Y offsets of a pair of alignment marks (or chip "footprints") on a substrate from their theoretical/nominal position, derives ΔX and ΔY corrections, and from the two mark positions computes a θ rotation and a shrinkage factor E. Correction values are applied to the X, Y, θ and the spacing/periodicity of prelocated chips; a mechanical "batch shrink compensator" physically merges/spreads the row of chips before a vacuum pencil transfers them to the substrate.
• §102 assessment (claims 1–4): This is the closest reference, and the present patent's own specification discusses it and expressly distinguishes it (mechanical compensator requiring servos). Under §102 each claim must be anticipated by a single reference. US4342090 discloses the informational core of claims 1(f) and 1(g) — measuring mark separation, comparing to nominal, developing a scaling ("shrinkage") factor, and applying it to nominal placement values (see its EQU2 shrinkage factor and the equation "X_T = X_(nominal) × E"). It therefore potentially anticipates the interpolation-factor concept as it appears in claims 1, 2, 3 and 4, if a tribunal reads "interpolation factor" broadly to cover the shrinkage factor E. However, on the face of the record it does not disclose at least: (i) optical means associated with the robot arm (in US4342090 the optics are fixed and the X-Y table moves the substrate); and (ii) robot arm control means that applies the factor to each site's nominal data and controls a robot to pick and place (US4342090 corrects spacing mechanically via a batch compensator and moves chips as a batch). My assessment: strong §103 reference; likely not a complete §102 anticipation of issued claims 1–4.

2.2 US4628464A — "Robotic system for mounting electrical components"

• Citation: US 4,628,464 A; issued 1986-12-09; filed 1983-10-07; Westinghouse Electric Corp.
• Description: A robot (support 10, programmable arm 12) with a conventional X-Y positioning mechanism 16 carrying a circuit board 14; a first TV camera 18 views the board position markers 15 and 17 to determine the board's orientation relative to the robot; a second camera 20 images each integrated circuit for orientation/centering. Data representing the location of each IC on the board in relation to the position markers is stored in the digital computer 17, which issues computer-generated move instructions; components are picked up with a vacuum probe 28 and placed on pads.
• §102 assessment: Discloses elements (a) robot arm, (b) optical means, (c) mounting station, (d) storage means with nominal component-site data, and (g) robot arm control means for pick-and-place. It does not appear to disclose measuring a distance between registration marks and comparing it to a nominal distance to derive an interpolation factor applied to each nominal site (element 1(f)/1(g) scaling). Its markers are used for orientation rather than for computing a board-wide scaling/interpolation factor. Therefore it does not anticipate claim 1, 2, 3 or 4 as issued, but is a pertinent §103 reference on the robot-arm/vision/stored-placement-data elements.

2.3 US4641257A — "Measurement method and apparatus for alignment"

• Citation: US 4,641,257 A; issued 1987-02-03; filed 1983-07-07; Canon Kabushiki Kaisha.
• Description: A step-and-repeat mask aligner (stepper). Alignment marks on a mask and a wafer are detected by an image pickup (TV) at a wafer-alignment station; a position detecting circuit with slice-level determination measures the pulse intervals between alignment-mark detection signals to derive relative position; the stage is moved by the measured deviation.
• §102 assessment: Discloses optical detection and measurement of alignment marks and a measured positional deviation — i.e., an anticipatory teaching for the "measure distance/compute correction" concept generally. But it is directed to photolithographic mask/wafer registration, with no robot arm, no component board, no component sites and no interpolation factor applied to nominal component-site data. It does not anticipate any of claims 1–4; it is at most a §103 (or background/analogous-art) reference on mark-based optical measurement.

2.4 US4723221A — "Method for manufacturing semiconductor device and apparatus therefor"

• Citation: US 4,723,221 A; issued 1988-02-02; filed 1983-11-26; Nippon Kogaku K.K.
• Description: Semiconductor-device manufacturing method and apparatus involving alignment/positioning of a substrate (wafer) for exposure/processing.
• §102 assessment: A semiconductor-fabrication alignment reference. On the available information it does not disclose a component-placement robot arm, component sites/footprints, or derivation/appliance of an interpolation factor to nominal component-site data, so it does not anticipate claims 1–4. (Caveat: I was unable to complete live verification of this reference's full text — my characterization rests on the patent's own citation and general knowledge, and should be treated as lower-confidence than §2.1–2.3.)

2.5 US4780617A — "Method for successive alignment of chip patterns on a substrate"

• Citation: US 4,780,617 A; issued 1988-10-25; filed 1984-08-09; Nippon Kogaku K.K.
• Description: Successive alignment of chip patterns on a substrate — a step-and-repeat alignment technique (wafer/reticle alignment in lithography).
• §102 assessment: Again a lithographic successive-alignment reference. No pick-and-place robot, no component-site interpolation. Does not anticipate claims 1–4; background/§103 only. (Same verification caveat as §2.4.)

2.6 US4720635A — "Automatic accurate alignment system"

• Citation: US 4,720,635 A; issued 1988-01-19; filed 1984-12-17; Disco Abrasive Systems, Ltd.
• Description: Wafer-cutting alignment system using key-pattern memory and pattern matching: an image-frame memory stores camera images of x-y matrix-arrayed pixels; first and second key patterns (before/after a 90° rotation) and their positions are stored; pattern matching plus coordinate transformation positions the wafer for cutting.
• §102 assessment: Discloses camera means, image memory and stored position data for alignment (relevant to storage means and optical means). It contains no component placement, no robot arm pick-and-place, and no interpolation factor derived from measured-vs-nominal registration-mark distance. Does not anticipate claims 1–4; analogous-art/§103 reference at most.

2.7 US4737845A — "Method of loading surface mounted device and an apparatus therefor"

• Citation: US 4,737,845 A; issued 1988-04-12; filed 1985-10-11; Hitachi, Ltd.
• Description: Method and apparatus for loading surface-mounted devices (SMDs) — i.e., surface-mount placement.
• §102 assessment: The most field-analogous citation (surface-mount component loading) and thus a strong §103 combination candidate with US4342090/US4628464. Its abstract/title do not evidence derivation of a measured-vs-nominal registration-mark interpolation factor applied per component site and generation of corrected site data for a robot to pick and place, so it does not anticipate claims 1–4 on the record. (Same verification caveat as §2.4 — full text unverified.)

3. Prior art discussed in the specification body (not in the 7-reference list)

The specification independently discusses two further references in the "Description" narrative. These are prior art relied on by the applicant for背景/patentability discussion even though they are not among the 7 listed citations:

3.1 US4681452 to Watanabe (as stated in the specification)

• Citation as stated: "U.S. Pat. No. 4,681,452 to Watanabe issued July 21, 1987"; described as an "automatic component inserter."
• Description (per specification): Displacement of insertion holes from nominal for a through-the-board leaded component is detected by light projected by the placement head and sensed below the holes; the actual board position is adjusted manually until maximum light passes, and the result is stored as a taught point for the X-Y positioning table. Requires repetition for every component (or several strategic ones); workable only for through-board components; tedious per board.
• §102 assessment: It is a manual taught-point approach and does not derive an interpolation factor from measured-vs-nominal mark distances applied per site; it does not anticipate claims 1–4.

3.2 US4651203 to Peterson (as stated in the specification)

• Citation as stated: "U.S. Pat. No. 4,651,203 issued May 17, 1987"; described as a "video controlled article positioning system."
• Description (per specification): Accurately places each component individually with video control — "extremely slow if applied to a printed circuit board where 300 to 400 components have to be placed."
• §102 assessment: Individual per-component visual alignment. It does not disclose a board-wide measured-vs-nominal registration-mark interpolation factor applied to nominal site data. Does not anticipate claims 1–4.

(Caveats: I could not complete live verification of the exact text of US4,681,452 and US4,651,203; the numbers, dates and descriptions above are reproduced literally from the authoritative patent text, per the instruction not to auto-correct identifiers.)

4. Bottom line

• No single cited reference appears to be a clean §102 anticipation of issued claims 1–4. Anticipation requires every element in one reference; none of the seven discloses the full combination of a robot arm whose arm-mounted optical means is driven to nominal registration-mark positions, an interpolation determination means comparing measured vs. nominal mark distance to develop an interpolation factor, and robot arm control means that applies that factor per component site to generate corrected site data and then picks up and places components there.
• US4342090A (Caccoma/IBM) is the most relevant reference and the closest thing to a §102 threat: it does measure mark separation, derive a shrinkage factor, and apply it to nominal placement values — i.e., it discloses the substance of the interpolation steps of claims 1(f)/1(g) (and thus touches claims 1–4 on that element). The specification's own distinguishing points are that Caccoma's correction is effected by a complicated mechanical "batch compensator" requiring servos, that its optics are not arm-mounted, and that it batch-corrects spacing rather than computing corrected data per site for a robot to pick and place.
• US4628464A (Westinghouse) and US4737845A (Hitachi) are the strongest §103 combination references for the robot-arm/vision/stored-placement-data and surface-mount-loading elements respectively.
• US4641257A, US4723221A, US4780617A and US4720635A are semiconductor alignment/registration references; they supply generic optical mark-detection and alignment teachings but do not anticipate the placement claims.

Confidence/limitations: Sections 2.1–2.3, 2.6 and 3 are grounded in my searches above and/or explicit specification text. Sections 2.4, 2.5, 2.7 and the full texts of §3.1–3.2 were not fully verified in this session (the search tool reached its step limit); those characterizations rest on the patent's own citation data plus general knowledge and should be re-verified against the full-text records (USPTO PatentCenter / Google Patents) before being relied upon. This is a technical analysis, not a legal conclusion on validity or infringement.

Key sources:

• https://patents.google.com/patent/US4922434/en (authoritative full text, claims, citations)
• https://uspto.report/patent/grant/[4922434](/patent/4922434)
• https://patents.google.com/patent/US4342090A (Caccoma/IBM)
• https://patents.google.com/patent/US4628464 (Westinghouse)
• https://patents.google.com/patent/US4641257A (Canon)
• https://patents.google.com/patent/US4720635 (Disco Abrasive)
• https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=[4342090A](/patent/4342090A) (and corresponding EP 0042977 A1)

I'll ground this analysis in the prior-art references listed on the patent page. Let me verify what each of the cited references actually discloses before mapping them to the claims.

Good — the key references are confirmed. Let me verify the remaining cited references and the two references discussed in the specification's Background but not listed in the front-page citation list.

Obviousness Analysis — US 4,922,434 (Fule, "Linear interpolation for a component placement robot")

Note on scope: My web-search budget was exhausted before I could open US4720635 (Disco), US4723221 (Nippon Kogaku), and the two references discussed only in the specification's Background. Those are explicitly flagged as unverified. Everything else below is grounded in the prior-art listing on the patent page plus the search results I retrieved, cited by URL. I have not auto-corrected any patent number or identifier.

1. Critical-date framework (this drives which references count)

Because §119 foreign priority cannot be used to extend the §102(b) one‑year grace period, the §102(b) critical date is 1987‑06‑16 (one year before the US filing). A reference published after 1987‑06‑16 but before 1988‑05‑25 is available only under §102(a); a US patent published after the filing date can still be §102(e) art as of its earlier US filing date. That distinction matters here (see the Nikon reference in §3).

2. Person of ordinary skill in the art (POSITA)

A POSITA at the 1988 priority date would be a robotics/automation engineer (BS ME/EE plus 2–4 years, or equivalent) with working familiarity with: (a) servo-driven pick‑and‑place robots and their controllers; (b) machine‑vision metrology on production equipment (fiducial/mark detection, pixel‑to‑inch scaling, camera calibration); (c) CAD/data‑driven placement file formats; and (d) the coordinate‑transformation mathematics routinely used in the semiconductor alignment/step‑and‑repeat art (scale, offset, rotation, orthogonality). This is a combination‑type skill set — mechanical placement plus vision metrology — which is itself a point in favor of obviousness (KSR, "a person of ordinary skill is also a person of ordinary creativity").

3. Prior-art inventory (from the patent page's own citation lists)

References in the page's "Cited By (20/27)" list that are NOT prior art to US4922434 (later priority dates): US5371836A (1988‑08‑24), US5084959A (1989‑09‑19), US5134665A (1990‑03‑08), WO1991019392A1 / US5321351A / EP0531428B1 (1990‑05‑25), and everything downstream. These may be used only as evidence of how the field evolved (and of the conventionality of vision-based fiducial correction), not as §102/§103 art.

4. Claim 1 — element mapping

Observation: claim 1 is a combination of (i) Caccoma's nominal‑vs‑measured mark comparison generating a scale/shrinkage correction, (ii) Westinghouse's marker‑relative, data‑driven robot place‑and‑pick control, and (iii) Nikon's algebraic packaging of the correction as a mapping applied to design data. Every element appears in the art; none of the claimed differences is a difference in kind.

5. Grounds of rejection under §103

Ground 1 (primary): Caccoma '090 in view of Westinghouse '464 — claims 1, 2, 3

What Caccoma alone supplies: the entire inventive insight the specification credits to the patentee — that artwork/board dimensional deviation can be quantified by sensing alignment marks relative to their nominal positions and reduced to a correction factor (ΔX, ΔY, shrinkage) applied to placement coordinates. Caccoma also supplies a board mounting station, optical sensing of marks, and storage of nominal data.

What Caccoma lacks: a robot arm that picks individual components and places them under software control, and the application of the factor to stored per‑site nominal data.

What Westinghouse supplies: precisely that — a programmable robot arm, cameras, and a computer holding "data representing the location of each integrated circuit on the circuit board in relationship to position markers", used to generate move commands per component.

Motivation to combine (KSR‑sufficient):

1. The specification itself supplies the motivation. US4922434 states: "Caccoma's 'batch compensator' is a complicated mechanical device requiring servos to allow merging and translation of rows of chips prior to their placement." That is the patentee's own articulation of a known deficiency in the closest art — an express "problem in the art" that a POSITA would be motivated to solve.
2. Same field, same problem, predictable substitution. Replacing a servo-driven mechanical merging/translation device with arithmetic performed on the coordinates already stored in a data‑driven robot controller is the substitution of one known technique (software coordinate scaling) for another (mechanical scaling), with predictable results. Caccoma already teaches the correction; Westinghouse already teaches the data‑driven per‑component application.
3. "Known technique" and "obvious to try." The patent expressly concedes that the hardware is off‑the‑shelf: "an off‑the‑shelf robot arm" and the AdeptOne robot with the Adept XGS Vision System, of Adept Technology, San Jose, Calif. A POSITA using such a standard robot/vision controller would inevitably have to express the correction as software operating on placement data, because that is how the controller works.
4. No new hardware, no new physics. The result (components placed at measured rather than nominal coordinates) was fully expected: the patent's own example shows the component would have been "underplaced by approximately 4 mils in the X direction and by 3 mils in the Y direction" — i.e., the correction performs exactly the arithmetic correction one would predict.

Ground 2: Caccoma '090 in view of Hitachi '845 — claims 1, 2, 3

Hitachi '845 teaches a robot that visually computes the displacement of position and rotation of the board contact pattern and of the component and controls the robot's operation in accordance with that information — i.e., vision-derived positional correction applied directly to robot placement ("relative position and orientation among them are also determined and surface mounted parts can be accurately loaded"). Combined with Caccoma's nominal‑vs‑measured scale factor, claim 1's (f)–(g) is squarely met, and the "why combine" rationale is the common field of robotic surface‑mount placement and the common objective of compensating board dimensional error.

⚠️ Counter‑point the patentee would raise (teaching away): Hitachi '845's Background criticizes exactly the architecture of claim 1(e) — using a dedicated mark on the board and moving the TV camera to it — as causing "an alteration of the design of the board and an increase in the time required for recognition due to the delay incurred by moving the TV camera." A patentee could argue this disparagement of "move the camera to the mark" teaches away from claim 1's "optical control means for … moving said robot arm … so that said optical means is positioned over the nominal position of consecutive ones of said registration marks." Rebuttal available: (i) Hitachi's complaint is about zooming onto tiny special marks and repeated camera moves — a throughput concern, not a technical impossibility, and it concerns marks "other than a contact pattern," whereas US4922434 uses marks that are part of the artwork itself; (ii) the same reference still teaches the operative principle (vision‑measured distortion → corrected robot placement), which is what claim 1 claims; (iii) Caccoma already teaches mark viewing. This is likely the patentee's best non‑obviousness argument, and it is a genuine factual dispute, not a slam dunk either way.

Ground 3: Caccoma '090 in view of Nikon '617 (secondary, §102(e) art) — claims 1, 2, 3

Nikon '617 is the closest art to claim 1's mathematics: it measures substrate marks, then determines linear expansion and contraction R (the interpolation factor), rotation θ, orthogonality, and two‑dimensional offset, and then builds a "corrected arrangement map of chips based on parameters and design data" — i.e., the nominal design coordinates are systematically corrected before positioning. That is claim 1(f)+(g) in algebraic form, applied to a distorted substrate carrying a nominal coordinate pattern.

Motivation: the problem statement is identical in substance — "satisfactory superposition accuracy has not always been obtained under the influence of the expansion and contraction of the wafer, the rotation error of the wafer on the stage, the degree of orthogonality of movement of the stage itself" (US4780617). The patent's own specification recites the same causal chain for boards: "the artwork may shrink or expand with respect to its nominal dimensions due to shrinkage or expansion of the board which carries it."

Anticipated patentee objection (non‑analogous art): semiconductor lithography (exposing resist on wafers) vs. robotic component placement on printed circuit boards. Rebuttal: the problem addressed — compensating nominal coordinate data for physical distortion of the workpiece — is "reasonably pertinent to the particular problem with which the inventor was involved," which is the controlling test. Nikon is best used as secondary support (particularly for claim 3's orthogonality/perpendicularity dimension), not as the lead reference, because of the field difference and the CIP/§102(e) date nuance.

Ground 4: any of Grounds 1–3 in view of Canon '4641257 — claim 4 (vision error correction factor)

Claim 4 adds: storage of a vision error correction factor, adjustment of the measured distance by that factor, then comparison to the nominal distance. The evidentiary posture is unusually favorable for a rejection:

1. The specification admits this is conventional. US4922434 states the vision-system readings "should be adjusted to account for this error and this may be done utilising standard calibration techniques", and that the error is found by measuring the same board with "any accurate measuring machine (MM) such as a Coordinate Measuring Machine or a laser inferometer." That is a §103 admission of conventional technique, independent of any reference.
2. Canon '4641257 discloses calibrating a camera-based mark-detection channel against an internal optical reference (mark position "determined on coordinates with its original of a TV wafer alignment reference mark provided within an optical system," with the spacing between that reference and the alignment position "precisely predetermined") and quantifies the residual detection error (±5 µm). This supplies the concept of characterizing and correcting a vision system's own measurement error before using its measurements for positioning.
3. KSR "design incentive" / "predictable variation": once one accepts (as Caccoma does) that placement coordinates must be corrected by a ratio of measured‑to‑nominal distances, a careful POSITA has an obvious incentive to eliminate known error sources from the measurement — and the accuracy of the robot's own vision system versus a coordinate measuring machine is exactly such a known, quantifiable, correctable source. The patent's example (X(MM)=15.2040 vs X(robot)=15.205; Y(MM)=10.6535 vs Y(robot)=10.655) is simple subtraction.

Ground 5: claims 2 and 3 — three marks, two perpendicular pairs

• Claim 2 (three registration marks): obvious as a design choice in view of Caccoma (a pair of marks with ΔX and ΔY offsets, per its abstract) combined with the need to span the board. More pointedly, the specification's final paragraph admits the substitute and the equivalence: "while the foregoing invention has been described in connection with the use of three artwork registration rings, clearly other artwork markings could be used, such as appropriate artwork footprints. Further, while … the nominal distance between the registration rings was stored …, if only the nominal co‑ordinates of these rings are stored therein, the nominal distance therebetween may be simply calculated." This is a self‑inflicted obviousness admission covering both claim 2's mark count and the data‑representation alternative of claim 1(d).
• Claim 3 (two pairs defining mutually perpendicular lines): a permissible variation/obvious design choice. Nikon '617's parameter set explicitly includes stage orthogonality and rotation error — i.e., the art already recognized that two non‑orthogonal measurement axes corrupt coordinate correction and that a second, crossed measurement direction is needed. Westinghouse's robot axes 34/36 and Caccoma's ΔX/ΔY sensing likewise make an X/Y (perpendicular) measurement pair a natural, expected implementation.
• Drafting caveat worth noting: the description renders the Y‑direction definition inconsistently ("a notional line through the centres of rings 25a and 25b define the (perpendicular) Y direction" — apparently a typo for 26a/26c, and it mixes tooling holes 25 with rings 26). Claim 3's "each pair defining a line perpendicular to the line defined by the other pair" should therefore be read on the specification's Figure 2 geometry rather than on the erroneous sentence.

6. Anticipation (for completeness, not the question asked)

Caccoma '090 alone does not anticipate claim 1: it lacks the robot arm of element (a) and the arm‑mounted optical means of (b), and its correction is embodied in a batch compensator rather than in control of a robot arm per element (g). The novelty of claim 1 therefore rests on the architecture, and the §103 case rests on the motivation to port Caccoma's arithmetic into the Westinghouse/Hitachi-style data‑driven robot — a case that, in my assessment, is stronger than not but is genuinely contestable at element (e) because of Hitachi '845's disparagement of camera-to-mark movement.

7. Confidence and unverified items

Bottom line: Claims 1–3 are, in my assessment, vulnerable to §103 rejections over Caccoma '090 + Westinghouse '464 (lead ground), reinforced by Hitachi '845 and Nikon '617. Claim 4 is additionally vulnerable because the patentee's own specification characterizes vision-system error calibration as "standard calibration techniques," and Canon '4641257 shows vision-based metrology being calibrated against a precise internal reference. The two areas that will decide the contest are (a) whether Hitachi '845's criticism of camera-to-mark movement rises to a teaching away from claim 1(e), and (b) whether the lithography alignment art (Nikon/Canon) is treated as analogous to component-placement art — I assess (b) as favorable to the challenger on a "pertinent to the problem" analysis, and (a) as a close call.