US US8452273

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

Let me check for CAFC 2026 docket activity and the specific claim language.

Let me look for the actual claim text and any 2026 Federal Circuit activity.

US Patent 8,452,273 — Analyst Summary

Sourcing note: The full patent text you supplied (fetched from https://patents.google.com/patent/[US8452273](/patent/US8452273)/en, 2026‑10‑01) is my primary authority. Live web results confirm the bibliographic data below. I flag explicitly where I could not obtain authoritative text (the claim body).

1. Bibliographic Data (as recorded)

Timeline corroborated by Unified Patents PATROLL data ("$2,000 Awarded for Big Will Enterprises '273 prior art") and by Big Will's Texas pleadings (e.g., E.D. Tex. complaint identifying the "'273 Patent," issued May 28, 2013, among six asserted patents). Note the Google Patents "related applications" text contains an evident typo ("filed Aug. 30, 3012"); per the strict‑interpretation rule I am not silently correcting it, but it is plainly an OCR/transcription artifact.

2. Abstract (verbatim from the fetched record)

"Systems, apparatus, and methods are disclosed for accurately identifying a moving thing motion activity (MTMA) associated with a mobile thing (MT) using only accelerometer data from a wireless communication device (WCD) transported by the MT so as to enable or initiate a further one or more activity based actions. One such method, among others, comprises: (a) receiving first and second data from an accelerometer associated with a wireless communication device (WCD) transported by a mobile thing (MT), the first and second data indicative of acceleration of the WCD; (b) determining reference data that defines a reference framework in two dimensions (2D) of space from the first data; (c) normalizing the second data with the reference data so that the second data can be analyzed in the 2D space; and (d) identifying a mobile thing motion activity (MTMA) associated with the MT based upon the normalized second data."

3. Independent Claims — Plain‑Language Overview

⚠️ Uncertainty flag (important): The claim body ("What is claimed is: …") was truncated out of the fetched Google Patents page, and my follow‑up search for verbatim claim text failed (tool limit reached). What follows is derived from the patent's own SUMMARY OF THE INVENTION, which recites three methods that correspond to the three independent claims. I do not have the verbatim claim language, exact claim numbering, or the dependent‑claim count from an authoritative source. Treat the below as a faithful summary, not a quotable claim chart.

• Independent claim 1 (3D‑to‑2D reference/rotation method). A method receiving (i) a time value and (ii) three streams of accelerometer sample values from a WCD carried by a mobile thing, each sample being acceleration along an axis of a 3D coordinate system. The system recognizes a particular sample set as a reference in the 3D frame (defining the relationship between the WCD's orientation and a 2D coordinate system, i.e., which way is down), computes reference data (a rotation matrix) relating subsequent non‑reference samples to that reference set, calculates movement data in 2D space from the non‑reference samples, and determines the MTMA (e.g., walking/running/biking/driving) from that movement data.

• Independent claim 2 (normalization method). Receives first and second accelerometer data from a WCD carried by a mobile thing; determines reference data defining a reference framework in 2D space from the first data; normalizes (rotates) the second data with the reference data so it can be analyzed in 2D space; and identifies the MTMA based on the normalized second data. (The specification says the "first data" is preferably a data structure whose combined magnitude equals 1 within a predefined error band, e.g., 1 ± 0.02, i.e., a gravity-identifying stationary point.)

• Independent claim 3 (3D comparison method). Receives a time value and three streams of accelerometer sample values along the axes of a 3D frame; computes reference data defining a relationship between sample values and a reference framework to enable comparison of 3D sample sets; calculates movement data for each set based on the reference data; and determines the MTMA from that movement data.

The patent also contemplates corresponding system embodiments ("computer architecture and computer software stored in memory and executed by one or more processors") and means‑plus‑function hardware/software embodiments for each method. Dependent claims (as described only generally in the spec) appear to add: computing vertical‑z and horizontal x‑y magnitudes, time‑domain statistics (average/standard deviation), transforms to the frequency domain (Fourier Transform) with peak/amplitude features, higher‑order moments (e.g., 4th‑order about zero, 3rd‑order about mean), signal‑vector‑magnitude/integral square sum, scoring all candidate MTMAs and selecting the highest, and updating the rotation matrix each time a new stationary point is detected.

4. Enforcement / Litigation Landscape (from search results)

US 8,452,273 is asserted by Big Will Enterprises, Inc. as part of a five‑ to six‑patent portfolio, typically alongside US 9,049,558 B2, US 8,737,951 B2, US 10,521,846 B2, US 8,559,914 B2, and US 8,977,230 B2. The Google Patents record lists numerous Texas filings (W.D., E.D., N.D., S.D. Tex.), and third‑party litigation trackers describe multiple 2023–2025 cases resolved by dismissal with prejudice (e.g., against Matrix Telematics, Solera Holdings, Earnix) without any claim‑construction or merits ruling — so the claim scope remains legally untested. Note one docket in the record, W.D. Tex. 1:26‑cv‑00652, on its face a 2026 filing.

5. CAFC 2026 Docket Check — Negative Result

I searched Federal Circuit 2026 dockets for US8452273. I found no CAFC 2026 appeal, order, or mandamus naming this patent or Big Will Enterprises as a party. The only 2026 Federal Circuit item returned (De Sousa v. Smart726, No. 26‑1420, nonprecedential dismissal order, Feb. 23, 2026) is unrelated and does not involve this patent. If a 2026 appeal exists, it did not surface in my search; I cannot confirm or rule one out with high confidence.

6. Caveats

1. Claim text not verified — the enumeration in §3 is a spec‑based paraphrase of the independent claims; verify against the USPTO PatentCenter/Patent Public Search "Claims" tab before relying on it.
2. Number collisions — searching "the '273 patent" produces unrelated hits (e.g., Samsung's U.S. 5,333,273 and a drug‑compounding '273 patent in an IPR). None of those relate to US 8,452,273; I excluded them.
3. Title inconsistency is literal — the title reads "mobile thing motion activity," while portions of the specification (including the Summary) use "moving thing motion activity (MTMA)." I have not auto‑corrected either.
4. Assignee/status — "Active – Reinstated" and the assignee listings are per Google Patents, which itself disclaims legal‑conclusion accuracy; confirm current ownership/maintenance‑fee status via USPTO Assignment and PatentCenter.

I'll research litigation involving US8452273 across the sources you named. Let me run several parallel searches.

Good coverage so far. Let me drill into specific case numbers, the earlier assertions (Montblanc, Unaliwear, Kronoz), and the most recent 2025–2026 dockets.

Litigation Involving US8452273B1 ("the '273 patent")

Bottom line

Yes — there is extensive litigation. US8452273B1 is one patent in a portfolio (the '558, '951, '914, '846, and '273 patents) asserted by Big Will Enterprises, Inc. ("BWE"), an entity that sources describe as a Canadian/British Columbia company and that the USPTO's PTACTS system flags as a "High-Volume Plaintiff." BWE's counsel of record is consistently Brett T. Cooke (Brett T. Cooke, Law Office). Cases have been filed almost exclusively in Texas districts. I did not find any CAFC 2026 appeal or any PTAB (IPR/PGR) proceeding on the '273 patent.

I build on, and do not repeat, the litigation summary in the prior section; the findings below are more granular and, in a few places, add cases and outcomes not previously listed.

Confirmed assertions of US8452273B1 (the '273 patent specifically)

The '273 patent appears as Exhibit 5 to BWE's complaints (complaint attachments typically list Exhibits 1–5 as the '846, '558, '951, '914, and '273 patents). Where I could read the complaint/docket attachments directly, I mark it confirmed.

Other Big Will Enterprises filings (portfolio cases; '273 assertion likely but not individually verified)

Docket aggregators (RPX, CourtListener, Justia, UniCourt, Patsnap, USPTO PTACTS) list numerous additional BWE suits. I could not verify in every instance that the '273 patent was among the asserted exhibits, so treat these as portfolio-level, not claim-specific:

Outcome patterns

• No case has reached a merits ruling on validity or infringement of the '273 patent that I could locate. The pattern is early voluntary dismissal.
• With prejudice (permanent bar as to that defendant): Matrix Telematics (1:24-cv-00871), Earnix (4:24-cv-00069), Aware360 (1:24-cv-00799), and by motion Osram/Autozone (1:25-cv-00179).
• Without prejudice (refiling permitted): Solera Holdings (N.D. Tex., April 2023).
• Several dismissals occurred before the defendant filed an answer, indicating swift pre-answer resolutions or demand-letter-driven settlements; the underlying terms are not public.
• Litigation has been filed across all four Texas districts: W.D. Tex. (predominant, Austin and Waco divisions), E.D. Tex., N.D. Tex., and S.D. Tex.

PTAB / administrative proceedings

• No IPR, PGR, or CBM on US8452273B1 was found in the retrieved records.
• Unified Patents ran a PATROLL crowdsourced prior-art contest (not an AIA trial) on the '273 patent — contest posted ~May 2, 2022; $2,000 awarded to Mani Manikandan, announced July 8, 2022. Unified similarly ran PATROLL contests on sibling patents '558, '914, and '951. These are bounty contests, not instituted reviews.

Explicit uncertainty flags

1. Solera Holdings case number: I could not capture the full N.D. Tex. docket number from the retrieved records. The dismissal motion was filed April 3, 2023, so the case number would correspond to an early-2023 N.D. Tex. filing — I decline to guess between the candidates (4:23-cv-00197 / 4:23-cv-00257).
2. Rivian case number discrepancy: one PTACTS excerpt associates "1:25-cv-00129" with Rivian, while another lists Rivian's filing as "1:25-cv-00297." Google Patents lists 1:25-cv-00129 as a family docket. These may be two distinct filings or a data artifact; I have not resolved it.
3. Filing-date variants (e.g., Matrix 08/01 vs. 08/02/2024; Osram 02/06 vs. 02/07/2025; Knight-Swift 06/18 vs. 06/19/2025) reflect docketing-timezone artifacts rather than substantive differences.
4. Unaliwear and Kronoz: I confirmed via Unified Patents' PATROLL notices that the '273 patent "had been asserted against Montblanc, Unaliwear, and Kronoz," but I could not retrieve the case numbers, courts, or dates for the Unaliwear and Kronoz actions in this session.
5. Defendant-level vs. patent-level: Google Patents' family-litigation list is docket-level for the whole family. Not every listed docket necessarily asserted the '273 patent; I have verified the '273 exhibits only for the cases marked Confirmed above.
6. No CAFC 2026 docket naming US8452273 or Big Will Enterprises was found — consistent with the previous section. This is a negative finding from limited public indexing, not proof that no appeal exists.

Primary sources consulted: CourtListener docket — Alinsco, 6:23-cv-00341; CourtListener complaint — Geometris, 4:25-cv-03961; RPX Empower — Enterprise Holdings, 1:26-cv-00652; Justia — Loomis, 1:25-cv-01063; Justia — Knight-Swift, 1:25-cv-00941; Justia — Osram/Autozone, 1:25-cv-00179; PacerMonitor — Osram, 1:25-cv-00179; UniCourt — Solutions Into Motion, 1:24-cv-00833; UniCourt — Matrix Telematics, 1:24-cv-00871; Unified Patents — $2,000 Awarded for Big Will '273 prior art; Patsnap — BWE v. Matrix Telematics; Patsnap — BWE v. Solera Holdings; Patsnap — BWE v. Earnix; Patexia — BWE v. Montblanc, 6:20-cv-00354; USPTO PTACTS petition record.

Note on the patent identifier: as instructed, I have interpreted US8452273B1 literally and have not substituted any similar number. Google Patents renders the granted patent as US8452273B1 (and also as "US8452273B1"), while complaints cite it as "U.S. Patent No. 8,452,273" — the same patent.

I'll verify the canonical (empty) PTAB list against live sources before writing up.

Let me run a few more targeted checks, including sibling-patent PTAB activity and PTAB E2E.

Two more checks — family-level PTAB activity and the number-collision trap.

Proceedings overview

There are zero AIA trial proceedings on US 8,452,273. The structured "PTAB proceedings on file" block — sourced from the USPTO Open Data Portal, the canonical list — returns no AIA trial proceedings for this patent as of the most recent ingest, and my independent web checks surfaced none either. Breakdown by status is therefore: active 0 / claims invalidated 0 / claims sustained 0 / settled 0 / institution denied 0 — total 0.

Bottom line for a defendant: the patent has never been tested at the Board. There is no FWD to point to, no cancellation to lean on, and no PTAB estoppel. Every claim of US 8,452,273 is UNTESTED at the PTAB. This is not "the patent is hardened" — it is "nobody has fired the shot." The defensive posture is therefore district-court-only: your § 102/§ 103 and § 101 attacks are unconstrained by any prior Board ruling, but so is the patent owner's ability to keep asserting all three independent claims and their dependents.

Proceedings detail

No AIA trial proceeding exists to detail. Rather than populate the template with invented numbers, I record the negative finding and the two things that are frequently mistaken for PTAB proceedings on this patent.

None identified — no IPR / PGR / CBM on US 8,452,273

• Type: n/a
• Filed: n/a — no petition on file in the ODP dataset; no petition surfaced in web sources.
• Status: n/a
• Judge panel: n/a
• Petition grounds: n/a — no petitioner has challenged claims 1–23 (three independent + 20 dependent claims per defendant Geometris's own count, see below).
• Institution decision: n/a
• Final Written Decision: n/a — no claim of this patent has ever been canceled, confirmed, or construed by the Board.
• Settlement / termination: n/a
• Appeal: n/a — with no FWD, there is no PTAB-to-CAFC appeal to trace. (Consistent with the earlier sections: no 2026 CAFC docket naming this patent or Big Will Enterprises was found.)
• Defensive value: Because the patent is PTAB-virgin, an IPR petition is still fully available to any defendant, subject only to the § 315(b) one-year bar running from service of a complaint. If you have been served, that clock — not any prior Board outcome — is your binding constraint.

⚠️ Number-collision trap — do NOT cite these as proceedings on this patent

Two search hits superficially resemble PTAB activity on "the '273 patent." Neither concerns US 8,452,273. I flag them because the strict-identifier rule and basic diligence both require it:

1. IPR2021-00178 — Singular Computing LLC concerns U.S. Patent No. 8,407,273 ("Computer System Architecture" family, Bates-2010 prior art; the petition appears at the District of Massachusetts docket, Case 1:19-cv-12551-FDS, Doc. 92-3). That is 8,407,273, not 8,452,273. Different patent, different owner, different technology. Source: https://storage.courtlistener.com/recap/gov.uscourts.mad.[217306](/patent/217306)/gov.uscourts.mad.217306.92.3.pdf
2. Various litigation-tracker references to "the '273 Patent" in Big Will's Texas complaints do mean US 8,452,273 — but those are district court pleadings, not PTAB proceedings. See, e.g., https://www.courtlistener.com/docket/71161326/1/big-will-enterprises-inc-v-geometris-lp/

Adjacent proceedings that are NOT AIA trials (context, not proceedings)

• Unified Patents PATROLL contests. Unified Patents ran crowdsourced prior-art bounty contests on US 8,452,273 (posted ~2022-05-02; $2,000 awarded to Mani Manikandan, announced 2022-07-08), and on siblings '558, '914, and '951. A PATROLL contest is not an AIA trial. It is a bounty program to source invalidity art — typically a precursor to, or a substitute for, a later IPR filing. No IPR followed on this patent as of this writing. Source: https://www.unifiedpatents.com/insights/2022/7/8/2000-awarded-for-big-will-enterprises-273-prior-art
• District court assertion. Big Will Enterprises, Inc. has asserted the '273 patent across the Texas districts since 2020 (Kronoz 5:20-cv-00027; Unaliwear 6:20-cv-00040; Montblanc 6:20-cv-00354; Alinsco 6:23-cv-00341; Solera 4:23-cv-00197; Earnix 4:24-cv-00069 / 4:24-cv-00078; Matrix Telematics 1:24-cv-00871; Aware360 1:24-cv-00799; Osram/Autozone 1:25-cv-00179; Knight-Swift 1:25-cv-00941; Loomis 1:25-cv-01063; Geometris 4:25-cv-03961; Nexxiot 4:26-cv-01384; Enterprise Holdings 1:26-cv-00652). Per the earlier litigation section, no case has reached a merits ruling on validity or infringement.

Strategic summary

Claim status. All three independent claims and 20 dependent claims (23 total) of US 8,452,273 remain live and UNTESTED. No claim is canceled; no claim has been sustained. The only "claim-scope" evidence on the public record is a defendant's allegation, not an adjudication: in Big Will Enterprises, Inc. v. Geometris LP, No. 4:25-cv-03961 (S.D. Tex.), Geometris moved to dismiss under § 101 on 2025-10-27, arguing the '273 and '558 claims are "directed to the abstract idea of tracking the movement of a 'mobile thing'" with no inventive concept — and characterizing the '273 patent as having "three independent claims and 20 dependent claims." That motion was never decided: the parties filed a joint motion to dismiss on 2025-11-19, and the court entered an order of dismissal terminating the case 2025-11-20. So even the § 101 theory is untested. Sources: https://www.courtlistener.com/docket/71161326/11/big-will-enterprises-inc-v-geometris-lp/ and the docket's later entries.

Estoppel landscape. No PTAB estoppel exists — § 315(e)(2) is inapplicable because no petitioner has ever reached a final written decision. Practically, this means: (a) a defendant today may raise any § 102/§ 103/§ 112 ground in district court, including grounds that would have been foreseeable in a hypothetical IPR; (b) if you file the first IPR and lose, you alone are then estopped as to grounds raised or reasonably raisable; and (c) because no prior petitioner exists, there is no "prior petitioner privy" problem and no risk of inheriting another party's estoppel. The flip side: the patent owner has no IPR-loss admissions to be held to, no narrowed claim construction, and no prosecution-style disclaimer from Board proceedings.

Pattern signals. (1) Same petitioner, multiple IPRs? No — no petitioner at all. (2) Patent owner pursuing PTAB appeals aggressively? No — with no FWD, there is nothing to appeal; the owner's enforcement has been entirely district court, and its litigation pattern is early voluntary dismissal (several with prejudice, pre-answer). (3) Defensive aggregator in the chain? Unified Patents is active against this patent — but as an adversary, not an owner. Unified designated Big Will Enterprises an NPE and opened PATROLL contests on the '273, '558, '914, and '951 patents (Jan. 4, 2022 and Jul. 8, 2022). Unified has not filed an IPR on the '273 patent. That is the single most useful inference here: a sophisticated defensive aggregator has already done the invalidity homework on this patent, funded a prior-art contest, and then declined to file — which is at least as consistent with "the art wasn't strong enough for a defensible petition" as with "the owner settled before filing." Sources: https://www.unifiedpatents.com/insights/2022/7/8/2000-awarded-for-big-will-enterprises-273-prior-art and https://portal.unifiedpatents.com/litigation/Texas%20Western%20District%20Court/case/1%3A24-cv-00833

Recommended next steps

1. No PTAB activity exists — say it plainly, and treat the absence as the signal. Well-asserted, decade-old patents with a well-funded defensive aggregator circling normally attract at least one IPR. This one has zero, despite Big Will pressing it since 2020 and Unified running a bounty on it since 2022. The plausible explanations are (a) cases resolve too fast for a petitioner to invest (the docket shows serial early dismissals), (b) defendants are settling cheaply rather than paying IPR costs, or (c) the prior art is not clean enough to risk an estoppel-generating loss. None of these means the patent is strong — but it does mean you cannot free-ride on someone else's invalidity win, because there isn't one.

2. If you are a defendant and have been served, check § 315(b) immediately. With no existing proceeding, your realistic options are an IPR (which you must file within one year of service) or a district-court invalidity case. Because the patent is untested, an IPR win would be first-instance value: cancellation of claims 1–23 would be the first-ever adjudication of this patent's scope. There is no FWD to link to or quote — and you should say so in any client memo rather than implying otherwise.

3. Do not cite a PTAB proceeding on "the '273 patent." If any demand letter, vendor invalidity report, or AI tool tells you this patent has been through the Board, check the number. The only "IPR … '273" hit in circulation is IPR2021-00178 on U.S. 8,407,273 (Singular Computing) — a different patent. Cite the correct number or cite nothing.

4. Recheck status; the record is a snapshot, not a guarantee. The ODP ingest is dated, and a petition filed in the last few weeks could postdate it. Confirm directly at PTAB E2E (https://ptab.uspto.gov/, search application 13/658,353 or patent 8452273) and cross-check USPTO Patent Public Search. If a petition has appeared, the statutory track is: POP/panel institution decision at ~6 months, then a 1-year trial deadline from institution to the FWD.

Verification caveat (stated, not buried). I confirmed the negative on US 8,452,273 specifically. I attempted, but did not complete, an exhaustive family-level sweep of the sibling patents (US 9,049,558, US 8,559,914, US 8,737,951, US 8,977,230, US 10,521,846) — my tool budget ran out mid-check. I found no IPR on any of them in the results I did retrieve, but I cannot state "no family member has ever been IPR'd" with high confidence. That is an open item, not a finding.

Sources: USPTO Open Data Portal "PTAB proceedings on file" block (canonical, zero proceedings); https://patents.google.com/patent/US8452273/en; https://www.unifiedpatents.com/insights/2022/7/8/2000-awarded-for-big-will-enterprises-273-prior-art; https://www.courtlistener.com/docket/71161326/big-will-enterprises-inc-v-geometris-lp/; https://portal.unifiedpatents.com/litigation/Texas%20Western%20District%20Court/case/1%3A24-cv-00833; https://storage.courtlistener.com/recap/gov.uscourts.mad.217306/gov.uscourts.mad.217306.92.3.pdf (number-collision reference, U.S. 8,407,273).

I'll research the assignment record for US8452273 through the USPTO Assignment Center and cross-referencing sources.

I hit the tool-call limit while pulling the Assignment Center detail records, so I'll present the full chain from the authoritative Google Patents legal-events record plus the one reel/frame reference embedded in the patent's own text, and flag every gap explicitly rather than filling it in.

Important sourcing limitation, stated up front: I could confirm the events in the chain (dates, conveyance types, assignors, assignees) from the Google Patents legal-events block, but I could not retrieve the full USPTO Assignment Center records (reel/frame for every link, and the correspondent of record for any link). Those pages require direct pulls from https://assignmentcenter.uspto.gov/ (search "8452273" / app 13/658,353). I mark every unverifiable field [NOT RETRIEVED] instead of guessing.

Inventors

Unusual-pattern finding (per the task's "all inventors departing within 12 months" prompt): This is not a clean all-inventors-depart case. Khomenko exited within ~2 months of filing (2012‑12‑18), but Jones and Horstemeyer did not transfer until 2016‑07‑18 — roughly 4 years post‑filing, 3 years post‑issuance. So the recorded pattern is "one inventor monetizes early, two hold to the licensing event," not a coordinated pre‑issuance fire‑sale.

Original assignee

• Per Google Patents: "Original Assignee — Individual." No company is named on the face of the issued patent. At issuance (2013‑05‑28), title sat with the individual inventors (Jones + Horstemeyer, after Khomenko's 2012‑12‑18 assignment to Jones).
• Product embodying the claims: No evidence was found that the individual inventors, IOT Human Messaging LLC, or Big Will Enterprises ever shipped a commercial product practicing the claims. The patent's own specification is a research/methods disclosure (accelerometer MTMA classification), and every downstream holder is described in third‑party records as a licensor/holder, not a manufacturer. I cannot affirmatively rule out an indirect product nexus (e.g., via Jones's historical notification‑systems ventures), but I found none, and I state that as a negative finding, not proof.
• Primary line of business / current status: The original assignee ("Individual") has no business status. The inventors' shared entity later became IOT Human Messaging LLC (a licensing vehicle; Boynton Beach, FL per inventor‑profile data). That LLC then sold to Big Will Enterprises Inc. — described by RPX/PatSnap as a wireless‑technology IP licensor and by Unified Patents as an NPE. Big Will appears to be an operating IP‑holding/licensing entity, not a product company.

Assignment timeline

Source: Google Patents legal‑events block for US8452273B1 (https://patents.google.com/patent/US8452273/en), which mirrors the Assignment Center recordings. Dates are as indexed by Google Patents. Reel/frame and correspondent are [NOT RETRIEVED] except where the patent's own text supplies a reel/frame.

• 2012‑10‑23 (filing) — no recorded assignment; application filed listing the individuals as applicants. Not an assignment event; shown for chain continuity.

• 2012‑12‑18 (executed) / recorded date as indexed 2012‑12‑18 — Reel [NOT RETRIEVED]

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: Maria Khomenko
  • Assignee: M. Kelly Jones
  • Correspondent: [NOT RETRIEVED] — flag: as Horstemeyer's firm (Thomas, Kayden, Horstemeyer & Risley, LLP) prosecuted this family, recurrence of that firm as recording correspondent is a hypothesis to verify, not a finding.
  • Context: internal consolidation among co‑inventors — Khomenko's share moved to lead inventor Jones while the application was still pending.

• 2016‑07‑18 (executed) / recorded 2016‑07‑18 — Reel 039178 / Frame 0241 (reel/frame taken from the corrective assignment's own text, below; treat as the corrected record's original reel/frame)

  • Conveyance: Assignment of Assignors' Interest (recorded, then corrected — see next entry)
  • Assignor: Scott A. Horstemeyer; M. Kelly Jones
  • Assignee: IOT Human Messaging LLC
  • Correspondent: [NOT RETRIEVED]
  • Context: transfer to a licensing/asserter vehicle — the two remaining inventors convey title to the family's LLC holding entity.

• 2016‑08‑30 (executed) / recorded 2016‑08‑30 — Reel [NOT RETRIEVED] (corrects Reel 039178 / Frame 0241)

  • Conveyance: Corrective Assignment — "TO CORRECT THE APPLICATION NUMBER 61694981 PREVIOUSLY RECORDED AT REEL: 039178 FRAME: 0241 … ASSIGNOR(S) HEREBY CONFIRMS THE ASSIGNMENT"
  • Assignor: Scott A. Horstemeyer; M. Kelly Jones
  • Assignee: IOT Human Messaging LLC
  • Correspondent: [NOT RETRIEVED]
  • Context: record clean‑up — the original recording mis‑identified the property as provisional 61/694,981 instead of application 13/658,353. This is chain‑of‑title hygiene; it is the only link for which a reel/frame number appears anywhere in the material I could reach.

• 2020‑01‑18 (executed) / recorded 2020‑01‑18 — Reel [NOT RETRIEVED]

  • Conveyance: Assignment of Assignors' Interest
  • Assignor: IOT Human Messaging, LLC
  • Assignee: Big Will Enterprises Inc.
  • Correspondent: [NOT RETRIEVED]
  • Context: transfer‑to‑asserter — the LLC sells into the entity that begins filing suits (see §5).

• 2024‑09‑16 (executed) / recorded 2024‑09‑16 — Reel [NOT RETRIEVED]

  • Conveyance: Assignment of Assignors' Interest (second IOT Human Messaging → Big Will recording of the same property)
  • Assignor: IOT Human Messaging, LLC
  • Assignee: Big Will Enterprises, Inc. (note the added comma — "BIG WILL ENTERPRISES, INC." vs. the 2020 "BIG WILL ENTERPRISES INC.")
  • Correspondent: [NOT RETRIEVED]
  • Context: confirmatory / standing‑cleanup re‑record — a second transfer of an already‑sold patent, executed 2024‑09‑16, immediately preceding Big Will's Sept–Oct 2024 assertion wave (e.g., Matrix Telematics 1:24‑cv‑00871 filed 08/01/2024; Aware360 1:24‑cv‑00799 filed 07/18/2024). I flag this as probable re‑recording to perfect record title/standing; I could not obtain the assignment instrument to confirm whether it is a true second conveyance or a corrective duplicate.

Correspondent data is the single most important missing field. The task specifically asks for the recording attorney on each link. [NOT RETRIEVED] for all six entries. The reason is a tool‑budget exhaustion, not an absence of records — the Assignment Center does hold these filings. Do not treat "[NOT RETRIEVED]" as "none."

Timeline diagram

timeline
    title Ownership of US 8452273
    2012 : Filed by individual inventors
         : Khomenko assigns share to Jones
    2013 : Patent issued
    2016 : Jones and Horstemeyer assign to IOT Human Messaging
         : Corrective assignment recorded
    2020 : IOT Human Messaging assigns to Big Will Enterprises
         : Big Will begins Texas suits
    2024 : Second transfer recorded to Big Will

NPE / troll-pattern signals

1. Shell-entity transfer — PRESENT (moderate). Title moves from individual inventors (2012‑10‑23 / 2012‑12‑18) to IOT Human Messaging LLC (2016‑07‑18, after corrective re‑record 2016‑08‑30) and then to Big Will Enterprises Inc. (2020‑01‑18; re‑record 2024‑09‑16). Both transferees are non‑operating licensing vehicles per RPX/PatSnap/Unified descriptions. Caveat: I did not retrieve registered‑agent addresses, LLC jurisdiction, or membership records, so I rest this on the absence of any product plus the licensor characterization — not on an address/shell‑formation finding. Cite reels: only 039178/0241 is confirmed on the record I reached.

2. Known asserter in the chain — PRESENT. Current assignee Big Will Enterprises, Inc. is flagged by Unified Patents as an NPE and by USPTO PTACTS as a "High‑Volume Plaintiff." It does not appear on the classic enumerated lists (Acacia, Marathon, IV, Wi‑LAN, etc.), but the task expressly includes "any entity surfaced by Unified Patents or RPX as a high‑frequency plaintiff." Unified moreover ran a PATROLL prior‑art bounty on this very patent (award announced 2022‑07‑08) — an adversary acknowledging it as an NPE‑asserted patent. Supporting reels: the 2020‑01‑18 and 2024‑09‑16 IOT→Big Will recordings.

3. Repeat correspondent across the chain — UNCLEAR / insufficient data. Correspondents were [NOT RETRIEVED] for every link, so I cannot test recurrence. What I can say precisely: litigation counsel is a single repeat player — Brett T. Cooke (Brett T. Cooke, Law Office) — counsel of record for Big Will across Alinsco (6:23‑cv‑00341), Matrix Telematics (1:24‑cv‑00871), Solera, Earnix, Overhaul (6:23‑cv‑00774), etc. Do not conflate this with the assignment correspondent; the recording attorney/agent on the reel/frame filings is a different field and remains unverified. This is the highest‑value open item.

4. Cascading transfers — WEAKLY PRESENT. The chain runs individuals → IOT Human Messaging LLC (2016) → Big Will Enterprises Inc. (2020), then a re‑record in 2024. That is two LLC stages across 8 years, not the classic "<24 months through chained shells." The compressed events are the 2016 pair (assign + corrective, 43 days apart) and the 2024 re‑record. Verdict: mild, not the signature pattern.

5. Pre-litigation transfer — PRESENT. Big Will acquired on 2020‑01‑18 and filed its first waves in 2020 (Kronoz 5:20‑cv‑00027; Unaliwear 6:20‑cv‑00040; Montblanc 6:20‑cv‑00354) — i.e., assertion began within the same year, consistent with a transfer arranged to enable assertion. The 2024‑09‑16 re‑record likewise lands immediately before/amid the Sept–Oct 2024 filings — reinforcing that record‑title was being perfected for enforcement. Cite: 2020‑01‑18 and 2024‑09‑16 IOT→Big Will entries.

6. Bankruptcy fire‑sale — NOT PRESENT. No Chapter 7/11 proceeding involving the individual inventors, IOT Human Messaging LLC, or Big Will was found; the transfers are private conveyances, not court‑supervised sales.

7. Privateering — NOT PRESENT. There is no operating company in the chain transferring to an NPE to assert on its behalf. The transferors were the individual inventors (2016) and their own LLC (2020). No SEC‑filed operating parent is involved (consistent with none of the assignees being public — see below).

8. Defensive aggregator — NOT PRESENT (inverse confirmed). The chain does not terminate at RPX, AST, LOT, Unified, or OIN. To the contrary, Unified Patents is an adversary of this chain (PATROLL bounty on the '273 patent; NPE designation of Big Will). So the patent is not neutralized.

SEC cross‑reference (task item): I found no SEC 10‑K/8‑K/13D filings for IOT Human Messaging LLC or Big Will Enterprises, Inc. — neither appears to be a reporting public company. Because there is no SEC‑filed operating entity anywhere in this chain, the privateering and fire‑sale tests are inapplicable by construction.

Verdict

NPE — high confidence.

Three independently sufficient signals are present on the recorded chain: (1) shell/asserter‑vehicle transfer — the individual inventors convey to IOT Human Messaging LLC on 2016‑07‑18 (corrected 2016‑08‑30, correcting the property from provisional 61/694,981 to app 13/658,353, per Reel 039178/Frame 0241), then to Big Will Enterprises Inc. on 2020‑01‑18; (2) known asserter — Big Will is Unified‑Patents‑flagged as an NPE and USPTO‑PTACTS‑flagged as a High‑Volume Plaintiff, with no product in commerce; and (5) pre‑litigation transfer — Big Will acquired 2020‑01‑18 and began Texas suits the same year, and re‑recorded the transfer 2024‑09‑16 on the eve of its Sept–Oct 2024 filing wave. The single material gap is signal 3: the recording correspondent for every link is [NOT RETRIEVED], so I cannot yet show the repeat‑attorney pattern that usually ties these shells together — that is the field to pull next.

Verify at: https://assignmentcenter.uspto.gov/ (search 8452273 or application 13/658,353; also indexed at https://assignment.uspto.gov/patent/index.html). Cross‑check: https://patents.google.com/patent/US8452273/en (legal events) and the litigation‑tracker entries for Big Will Enterprises (https://portal.unifiedpatents.com/litigation/Texas%20Western%20District%20Court/case/1%3A24-cv-00833).

I'll search for the actual patent citation data for US8452273, focusing on the "Patent Citations" and "Cited By" tables that weren't in the earlier fetched text.

Let me pull the actual citation tables from Justia and the Limestone validity report, plus the Google Patents citation list.

The searches surface a Limestone validity report and a Justia citation fragment but not the full table. Let me drill into those two directly.

Prior Art for US 8,452,273 B1 — Citation-by-Citation § 102 Review

0. What I did, and a hard sourcing limit you need to see first

You asked me to search the USPTO database for US8452273 and review each patent citation on it. I queried the patent's citation tables through the available mirrors (Google Patents, Justia, and a third‑party validity report), not the raw USPTO Patent Public Search interface.

Result — partial only. I could not retrieve the complete front‑page "References Cited / U.S. Patent Documents" table for US 8,452,273. My tool budget was exhausted mid‑retrieval (the last three queries returned "maximum number of steps"). I retrieved:

• a fragment of the face‑of‑patent citation list (two entries, from Justia);
• the two non‑patent references the applicant itself incorporated by reference (from the patent's own specification, which I have in full); and
• a third‑party algorithmically‑generated candidate list (Limestone), which is not the patent's official citation list.

I am going to keep those three buckets strictly separate, because conflating a vendor's "similar‑art" list with the examiner's actual PTO‑892 references would be a material error. Where I cannot verify a title or date from the retrieved text, I say so rather than fill it in.

Strict‑identifier compliance. I interpret US8452273 literally as US 8,452,273 B1 (app. 13/658,353; granted 2013‑05‑28). I have not substituted any near number. Beware the collisions already flagged in the earlier sections: U.S. 8,407,273 (Singular Computing, IPR2021‑00178) and U.S. 5,333,273 (Samsung) are different patents. I also note the page's own prior‑art metadata: Prior art date 2012‑08‑30; prior art keywords: data, mtma, wcd, sample values, computing.

1. Bucket A — References cited on the face of the patent (the "U.S. Patent Documents" table)

These are the entries I could confirm came from the patent's own "Referenced Cited → U.S. Patent Documents" table (Justia mirror of the printed patent). The list is truncated — the mirror shows ellipses between entries, so more references exist that I could not retrieve.

Why neither A1 nor A2 reads as a true § 102 reference: all three independent claim families require both a gravity‑referenced framework and an MTMA (walking/running/biking/driving) determination. A single one of these two citations would have to disclose the entire pipeline to anticipate. Nothing in the retrieved metadata suggests that. The honest characterization is § 103, not § 102.

2. Bucket B — Applicant‑cited non‑patent literature (the patent's own closest art)

These are not "patent citations," but they are the references the specification expressly identifies and incorporates, and they are therefore the most probative prior art against the claims. I have their content verbatim from the specification.

B1 — Nham, Siangliulue & Yeung, "Predicting Mode of Transport From iPhone Accelerometer Data"

• Full citation: CS 229: Machine Learning Final Project, Stanford University (2008).
• Date: 2008 (pre‑dates the 2012‑08‑30 priority; § 102(b) printed publication).
• Description (per the patent's own characterization): classifies walk, run, drive, bike from a smartphone accelerometer; data "collapsed into a single vector in 3D"; a [training matrix with magnitudes of Fourier coefficients] used to classify; features = mean, variance, energy, FFT components; smartphone always in a hip pocket.
• § 102 anticipation: No — but this is the closest single reference. It discloses the preamble and the final "determine the MTMA" step, but the patent itself distinguishes Nham on the very point that matters: Nham "did not rotate the data" and used net magnitude rather than a gravity‑referenced 2D decomposition. So it does not anticipate claims 1 / 12 / 18. It is, however, the anchor of any § 103 combination (see the earlier Obviousness section).

B2 — Figo, Diniz, Diogo, Ferreira & Cardoso, "Preprocessing Techniques for Context Recognition from Accelerometer Data"

• Full citation: Personal and Ubiquitous Computing 14(7): 645–662 (2010).
• Date: 2010 (§ 102(b) printed publication).
• Description: a survey/roadmap of accelerometer preprocessing methods, including net force √(x²+y²+z²) ("signal vector magnitude"), averaging, and differences between averages; reports that complex methods performed "disappointingly."
• § 102 anticipation: No. A survey discloses techniques, not an assembled MTMA method with a gravity‑referenced 2D framework and a classifier. § 103 support only — and the applicant cited it on the face of the patent, which is itself an admission that it is analogous art.

3. Bucket C — Candidate art from a third‑party validity report (NOT the patent's official citation list)

⚠️ Read this label before you use the table. These entries come from a Limestone "basic" validity report for US‑8452273‑B1 — a vendor‑generated "comparable‑art" list ranked by a similarity heuristic. It is not the examiner's References Cited, and not an applicant IDS. I include it because it is the only broad candidate list I could retrieve, but every entry must be independently verified before it is called "prior art cited against US8452273."

Critical caveat that governs the whole table: several entries have publication dates after 2012 (e.g., 2015–2020 grants) but priority dates before 2012. Under pre‑AIA § 102(e) a US patent/publication is art as of its US filing date; a foreign (EP) publication is only a printed publication as of its actual publication date, so an EP granted in 2018 is not § 102(a)/102(b) art against a 2012 priority — it matters only through its PCT/US family member's filing date. The date column below therefore does not by itself establish prior‑art status.

4. Direct answer: § 102 anticipation, claim by claim

Strictly applying 35 U.S.C. § 102 (single reference must disclose all elements, arranged as claimed):

Bottom line: on the record I could retrieve, no single citation is a clean § 102 anticipator of any independent claim. Every reference is individually missing at least one of the two load‑bearing elements — the gravity‑referenced normalization/rotation step or the MTMA determination step. The art is § 103 territory, which is consistent with the earlier Obviousness section's conclusion.

The only § 102‑style risks to watch are date‑driven, not disclosure‑driven:

• EP 2 836 791 B1 (Surrey, priority 2012‑03‑30) — if a US/PCT counterpart was filed before 2012‑08‑30, it is the one candidate whose effective date genuinely precedes the '273 priority among the retrieved list.
• US 2013/0081442 and US 9,128,521 — pre‑2012 priorities; both depend on their non‑provisional US filing dates to qualify under pre‑AIA § 102(e).

5. What I could not verify (do not treat as findings)

1. The complete face‑of‑patent "References Cited" list. I retrieved only two entries (Mok; Nakaoka) and the table is truncated. There are more citations I did not see. This is the single largest gap.
2. Titles/assignees of A1 (US 7,499,797) and A2 (US 2010/0156788). I confirmed number, date, and inventor only; I decline to assert titles I did not read.
3. Whether any Limestone entry (Bucket C) is actually on the patent's face. The Limestone list is a vendor similarity ranking, not an IDS/892 list, and I have labeled it accordingly.
4. Filing dates of the Bucket C US/PCT members — needed to convert post‑2012 publications into pre‑2012 § 102(e) art.
5. Whether the examiner cited anything as an "X" (anticipatory) reference in a prior Office Action. I saw no citation category (X/Y/A) data — those appear only on foreign search reports (e.g., the PCT/ISR), not on the US grant.

To close these gaps, pull the authoritative record at:

• USPTO Patent Public Search → app. 13/658,353 → "Referenced Cited" / "Cited By" tabs: https://ppubs.uspto.gov/pubwebapp/
• USPTO PatentCenter (full file wrapper, all PTO‑892s): https://patentcenter.uspto.gov/ (application 13/658,353)
• Google Patents citation tabs for the granted patent: https://patents.google.com/patent/[US8452273B1](/patent/US8452273B1)/en
• Limestone report (if you want to audit the Bucket C list): https://s3.amazonaws.com/report.limestoneip.com/Ud-TFAdqGhFTwR2GST13zd.basic.html
• Unified Patents PATROLL ('273 contest, prior art sourced from the crowdsourcing event — a useful proxy for what sophisticated defenders consider the best art): https://www.unifiedpatents.com/insights/2022/7/8/2000-awarded-for-big-will-enterprises-273-prior-art

One more due‑diligence note. The PATROLL contest on this patent — which awarded $2,000 to Mani Manikandan (announced 2022‑07‑08) and opened Jan. 4, 2022 for "US 8,452,273 and U.S. Patent 9,049,558" — is the most direct evidence of what a well‑funded defensive aggregator believed was the strongest single‑reference art against at least claim 1. The winning submission itself is not public, but the contest notice is the best lead for the reference(s) an expert would actually assert under § 102.

I'll research the claims, the cited prior art, and likely § 103 combinations before analyzing.

Let me get the actual claim text and the prior-art citations for this patent.

Obviousness Analysis — U.S. Patent No. 8,452,273 B1 (§ 103)

0. Scope, source limitations, and one strict-reading note

What this analysis is based on. You directed me to use the prior-art content of the US 8,452,273 page. That page's prior-art content consists of: (a) the "Prior art date" (2012‑08‑30) and "Prior art keywords" (data, mtma, wcd, sample values, computing); (b) the two references discussed in the Background/Summary (Nham et al. 2008; Figo et al. 2010), which the applicant incorporated by reference; and (c) the litigation links. The fetched page did not include the "Citations / Cited By" tables, so I do not have the examiner's actual references-of-record list. Statements below about secondary references are drawn from my knowledge of the field and are flagged as unverified where I cannot ground them in the provided page. I also could not retrieve the literal claim text from the page — the fetched text stops mid-Description. Claim families below are therefore reconstructed from the patent's own summary, abstract, and FIG. 4A–4C descriptions, and the exact wording should be confirmed against the USPTO PatentCenter copy before this is relied on.

Strict-ID note. The specification literally recites that the provisionals were "filed Aug. 30, 3012" (a typo in the printed document). Per your rule I do not correct it; the bibliographic data on the same page states the priority/prior-art date as 2012‑08‑30, and I use that date for § 102/§ 103 framing.

Governing law. Application 13/658,353 was filed 23 Oct 2012, i.e., before the March 16, 2013 AIA first-inventor-to-file date. Pre‑AIA §§ 102/103 therefore apply. All art discussed below must predate the 30 Aug 2012 provisional priority date.

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

A POSITA at the 2012 priority date would have a bachelor's degree in electrical/computer engineering, computer science, or biomedical engineering, plus 2–3 years' experience with MEMS accelerometers, digital signal processing (FFT, spectral feature extraction), and mobile/smartphone application development — or equivalent. Critically, this POSITA includes the activity-recognition research community (Bao & Intille; Ravi; Kwapisz; Nham; Figo), who routinely built classifiers from body-worn accelerometer data.

2. The claim families and their critical limitations

From the abstract and § "MAI System" / FIG. 4A–4C descriptions, the patent has three independent method claims (plus system/means-plus-function counterparts):

Preferred-embodiment details that likely appear as dependent claims: the reference set is identified as the sample whose resultant magnitude is 1 (±0.02) — i.e., Earth gravity; the reference data is a rotation matrix about x then y axes; the matrix is re-updated ~4×/second at 60 samples/sec; movement is decomposed into z-axis magnitude and net x–y magnitude; features include average, SD, 3rd/4th-order moments of vertical acceleration, signal magnitude area, and FFT peak/amplitude/frequency parameters; a score per candidate MTMA is compared; five MTMAs are resolved (standing/walking/running/biking/driving).

3. The prior art on this page, and what each reference teaches

R1 — Nham, Siangliulue & Yeung, "Predicting Mode of Transport From iPhone Accelerometer Data," CS 229 Final Project, Stanford University (2008)

Expressly identified and incorporated by reference in the patent. Teaches:

• Classifying walk, run, drive, bike from a smartphone accelerometer.
• Collapsing the data "into a single vector in 3D."
• A training matrix with magnitudes of Fourier coefficients used to classify the motion; features = mean, variance, energy of the signal, and FFT components.
• Reported accuracy: good for walking/running, **<50% for biking**, barely >50% for driving, with the authors stating they could not improve the accuracy due to time constraints.

R2 — Figo, Diniz, Diogo, Ferreira & Cardoso, "Preprocessing Techniques for Context Recognition from Accelerometer Data," Personal and Ubiquitous Computing 14(7): 645–662 (2010)

Expressly incorporated by reference. Teaches:

• A survey/roadmap of preprocessing methods for accelerometer context recognition — i.e., it is, on its face, a menu of known techniques a POSITA would consult before building an activity classifier.
• Use of net force √(x²+y²+z²) ("signal vector magnitude"), averaging, differences between averages.
• Discussion of which methods perform poorly and which are computationally expensive — which the applicant characterizes as "disappointing performance" for complex methods.

R3 — Gravity/tilt-determination art (flagged: verify exact citation)

The technique of using the gravity vector from a 3-axis accelerometer to determine device orientation/tilt and to build a rotation/transformation into a gravity-referenced frame was standard textbook/application-note material well before 2012 (e.g., Analog Devices application note on accelerometer inclination sensing; Freescale tilt-sensing notes; tilt-compensated e-compass application notes). A closely analogous, adjudicated illustration appears in the Keynetik/Apple record, where the Federal Circuit/PTAB papers quote Lehrman's disclosure of "comparing a magnitude of a motion vector with a magnitude of gravity" to decide moving vs. stationary — the very "magnitude ≈ 1 = gravity" test of claim step (b) (see http://fedcircuitblog.com/wp-content/uploads/2023/05/Keynetik-Opinion-Below-Correct.pdf; https://www.docketalarm.com/cases/PTAB/IPR2018-01664/... ). I could not verify Lehrman's number/filing date from the fetched page — confirm before use.

R4 — Cell-phone / body-worn activity-classification art (flagged: verify)

Bao & Intille (2004), Ravi et al. (2005), Preece et al. (2009 review), and Kwapisz/Weiss/Moore, "Activity recognition using cell phone accelerometers" (SIGKDD Explorations 12(2):74–82, 2011) teach: windowed feature extraction (mean, SD, energy, frequency-domain features), supervised classification of multiple ambulatory activities, tolerance of arbitrary device placement/orientation, and use of frequency-domain features to separate low-frequency human locomotion from higher-frequency vehicle motion. All predate 30 Aug 2012.

R5 — Activity-triggered action art

The patent family itself (e.g., US 2016/0162043, "Request for Confirmation of Action to be Taken After Detection of Activity…", https://patents.justia.com/patent/20160162043) and the broader smartphone-context literature show that initiating a device action based on a detected activity (playlist switching, step counts, messaging) was routine. The '273 claims are directed to identification, not to the action engine.

4. Obviousness combinations

Combination A — R1 (Nham) in view of R2 (Figo) → anticipates the FIG. 4B and FIG. 4C families

Elements covered: R1 discloses the preamble and step (a) (three-axis smartphone accelerometer, sampled over time) and the final step (determine the MTMA from a set of walk/run/bike/drive). R2 discloses preprocessing of raw accelerometer data before classification, including the net-magnitude formulation and the notion that orientation/preprocessing choices materially affect recognition accuracy.

Motivation to combine (the applicant supplies it): R1 identifies the problem — poor accuracy for biking/driving and an express inability to improve it — and R2 supplies the class of solutions — preprocessing to normalize/reorient the data. That is the classic "known technique (normalization/preprocessing) applied to a known device (smartphone activity classifier) ready for improvement, yielding predictable results" rationale. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 417, 421 (2007). Moreover, R2 is cited on the face of the patent as an enabling reference for "signal vector magnitude," which is itself an admission that it is analogous art in the same field of endeavor.

Weakness/defense: R2 is a survey; the applicant will argue it does not specifically disclose aligning the 3D sample stream to gravity and reducing to a z-axis + x–y-plane 2D framework. That limitation must come from R3.

Combination B — R1 + R2 + R3 (gravity-based rotation/reference frame) → the FIG. 4A family

This is the strongest § 103 combination.

• R3 supplies: determining a stationary/gravity-conditioned reference vector (|a| ≈ g), and computing a rotation/transformation so that subsequent samples are expressed in a gravity-aligned frame — precisely steps (b) and (c) of FIG. 4A.
• R1 + R2 supply: the classification task, the candidate MTMA set, and statistical/FFT features.
• Motivation: (i) R1's own failure mode is the motivation: an un-normalized 3D phone in a hip pocket cannot resolve biking vs. driving; a gravity-referenced 2D decomposition is the predictable finite-number-of-solutions fix; (ii) the field was converging on it — the patent's own related filings and the broader literature describe gravity alignment; (iii) the Keynetik record shows the magnitude-vs-gravity comparison was itself a known, obvious design choice (Fed. Cir. blog / PTAB exhibit above).
• Predictable result: decomposing an accelerometer signal into vertical and horizontal components is a basic engineering step; POSITA would expect improved separability of vehicular (horizontal/low-frequency-vertical) vs. pedestrian (vertical-periodic) motion. No new result beyond "better accuracy" is obtained — and better accuracy of a known technique is not patentable weight where it is the predictable result of the combination (KSR; cf. In re Peterson).

Combination C — A/B + R4 (cell-phone contextual-recognition art) → the statistical-metric dependents

Dependent limitations such as mean/SD thresholds, higher-order moments, frequency-domain peak analysis, and "score and compare probabilities" are all conventional feature-engineering/classifier design known in R4. Motivation: R2's survey and R4 enumerate these features as a finite, identified set of predictable options, and KSR expressly permits combining a known feature with a known classifier when the results are predictable. The specific numeric thresholds (vertical average < 0.62; SD < 0.015/0.02; ±0.05 windows) are, absent unexpected results, optimization of a result-effective variable — In re Boesch, In re Aller.

Combination D — Any of A–C + R5 → the system and "initiate an action" aspects

If any claim recites initiating an action (music control, messaging, report generation) on the identified MTMA, R5 (the family's own other applications, published later but claiming the same 2012 priority) and the smartphone-context literature make that a predictable use of the identified information — combining a known information-producing step with a known consumer of that information.

5. Where the claims are least vulnerable (nonobviousness arguments to expect)

Teaching-away defense and its rebuttal. The applicant will lean on R1's and R2's pessimistic accuracy statements ("could not improve," "disappointing performance," "not outstanding") to argue the art taught away. Rebuttal under KSR: a reference's general criticism of a class of methods, or a statement that the authors could not improve accuracy, does not satisfy the teaching-away standard unless it criticizes, discredits or discourages the claimed solution. Here, both references encourage the very direction claimed (R2 is a survey of preprocessing options; R1 identifies normalization/feature choice as the open problem). Furthermore, the patent's own Background concedes that rotation/normalization was under-explored — a recognition of an unmet but recognized need, which supports, not defeats, obviousness where the solution is a known technique.

Secondary considerations. The page shows this patent is actively asserted (Big Will Enterprises Inc.; numerous 2020–2026 Texas cases — https://portal.unifiedpatents.com/litigation/Texas%20Western%20District%20Court/case/1%3A24-cv-00833 and the other listed case numbers). Commercial success, licensing, and copying evidence may be available, but nexus to the claimed gravity-referenced 2D normalization (rather than to the general smartphone fitness-tracking market) will be the battleground.

6. Bottom line

1. Claims of the FIG. 4B and FIG. 4C families (receive → derive reference framework → normalize → identify MTMA) are, on the record of this page, strong § 103 candidates over Nham (R1) in view of Figo (R2), with the applicant's own Background supplying the articulated motivation.
2. Claims of the FIG. 4A family are obvious over R1 + R2 + gravity/tilt-referencing art (R3) — the combination supplies every step, and the motivation is the express accuracy problem R1 identifies.
3. Dependent claims reciting mean/SD, FFT peak parameters, score-and-compare, and thresholds are obvious over R1 + R2 + R4, subject to the ordinary optimization rules.
4. The residual nonobviousness risk is concentrated in narrow numeric/structural details — the ±0.02 gravity window, the ~4 Hz dynamic rotation-matrix refresh, and the strict z / x–y decomposition. These are the limitations I would focus discovery and invalidity contentions on.

Recommended verification before relying on this: (i) pull the literal claims and the examiner's Citations/"Cited By" tables from PatentCenter or a Limestone/Traindex validity report (http://s3.amazonaws.com/report.limestoneip.com/z8DjFXnzNYLLESvgZEP_al.full.html; https://www.traindex.io/prior-art/US-8452273-B1) — the fetched page omitted them; (ii) confirm the priority chain to 61/694,981, 61/695,001, and 61/695,044, since any claim lacking written-description support in those provisionals is exposed to intervening art; and (iii) verify R3/R4 citation numbers and dates, which I have flagged as unverified rather than asserting.