US 9635675

I'll cross-check the provided full text against external sources, then summarize.

Let me check for any litigation or 2026 Federal Circuit activity tied to this specific patent.

One more check on family/related records and any appellate activity.

US Patent 9,635,675 — Verification & Summary

Source of record used: the authoritative full text fetched from Google Patents (https://patents.google.com/patent/[US9635675](/patent/US9635675)/en), cross-checked with a second independent mirror of the pre‑grant publication (US20150249996A1) at patents-review.com, and with the WO sibling WO2014038111A1. Where external sources agreed with the supplied text, I treated the supplied text as authoritative.

⚠️ Identifier caution (per your literal-interpretation rule): the bare digit string "9635675" also resolves to an unrelated PCT publication WO1996035675A1 ("1‑fluor‑5,6,7,8‑tetrahydroisoquinoline derivatives… liquid crystal mixtures"). That is a different document and is not the patent you asked about. Everything below concerns US 9,635,675 B2 only.

Bibliographic data

Abstract (as issued)

Problems with communication interface selection control using a fluctuation control equation are that it is difficult for the control to address large environmental variations and that the control requires parameter adjustments according to environments, which incurs operating cost. A communication method includes observing communication environment conditions to generate communication environment data, generating fluctuation information by using the communication environment data as the data occurs, and allocating communication resources by performing attractor selection control on the basis of the fluctuation information. Generating fluctuation information comprises using a fluctuation equation to perform fluctuation equation control utilizing the fluctuation information as it occurs. The fluctuation information is not fluctuation terms that are independent random noise terms in the fluctuation equation but includes a fluctuation term reflecting the communication environment data.

Plain-language overview of the independent claims

The core inventive idea across all independent claims: instead of feeding artificial Gaussian noise into the attractor‑selection fluctuation equation, the noise term is harvested from measured communication-environment data itself. The specification's Eq. (7) is the concrete implementation: η′ₖ = Cₖ(BERₖ − μₖ), i.e. instantaneous BER minus its historical average μₖ, scaled by a fixed parameter Cₖ.

• Claim 1 — Method (radio interface selection).

  1. Generate communication environment data by observing environment conditions.
  2. Calculate a fluctuation term used in the fluctuation equation from that data as it occurs, and run the fluctuation equation calculation using that term. The calculation specifically comprises: (a) subtracting the average of the history of the environment data from the observed data (mean‑shift toward zero), and (b) multiplying by a given parameter to adjust the fluctuation's amplitude for adaptation to environmental variations.
  3. Allocate resources in a selecting unit based on the equation's output. The environment data is communication quality information at multiple radio interfaces of a terminal that can talk to multiple base stations, and allocation = picking the radio interface with the largest output value (i.e., argmax over mₖ).

• Claim 2 — Method (network domain/cluster selection). Same three-step architecture and the same two-part fluctuation‑term calculation (history‑average subtraction + amplitude parameter multiplication), but the environment data is inter‑domain control signals at a communication node of a network made of multiple domains, and allocation = changing the domain a node belongs to, based on the equation outputs and the volume of inter‑domain control signals passing through that node.

• Claim 4 — Device (radio interface selection). The apparatus counterpart of claim 1: a communication‑environment‑data generating device, a fluctuation information generating device (calculating the fluctuation term and performing the fluctuation equation calculation), and a selecting unit; the generating device includes a fluctuation equation control unit performing the same mean‑shift‑then‑scale operation; the selector picks the interface with the largest output.

• Claim 5 — Non‑transitory computer‑readable medium. A stored control program causing a computer to carry out the claim‑1 procedure, including the same mean‑shift + amplitude‑adjustment procedure for the fluctuation term.

• Claim 6 — Device (domain/cluster selection). The apparatus counterpart of claim 2, with the same fluctuation‑term computation and domain‑change selection logic.

  • Literal‑reading note: claim 6 as printed contains what appears to be a drafting error — "a communication environment data generating device for generating communication environment data generating device by observing communication environment conditions." I flag it rather than correcting it, per the no‑auto‑correction rule; it does not appear to be a search‑result artifact, as it appears in the supplied authoritative text.

• Claim 8 — Non‑transitory computer‑readable medium. The stored program counterpart of claim 2 (domain change based on equation outputs and inter‑domain control‑signal volume).

Dependent claims 3, 7, 9 add the two‑tier mechanism: the control device of each domain computes fluctuation‑equation information for nodes belonging to its own domain, and the control formula for the domain‑change process uses the ratio of control‑signal traffic to user traffic passing the node (so a node carrying heavy control‑signal traffic migrates to another domain).

Technical framing and advantage asserted

Attractor selection is described as a bio‑inspired paradigm where the system state m evolves as dm/dt = α·f(m) + η, with activity α ∈ [0,1] weighting a control structure with an attractor against a fluctuation/noise term. The stated problem in the prior art (including the inventor's own NPL 1, G. Motoyoshi, N. Wakamiya, M. Murata, WONS '12, and Patent Literature 1 = JP 2011‑155508, NTT) is that η is independent Gaussian noise, so a designer must pre‑tune the noise standard deviation to expected environmental‑variation levels. The asserted advantage is elimination of that design‑stage tuning and its operating cost, because the data‑derived noise term correlates with the control terms and therefore self‑scales with environmental variation.

Analytical observation (mine, not the patent's): the granted independent claims nonetheless recite multiplying by a "given parameter … to adjust an amplitude … for adaptation to environmental variations" (Cₖ in Eq. 7), so the claims are arguably narrower than the "no amplitude tuning needed" benefit recited in the Advantageous Effects section. Anyone analyzing scope or validity should treat that tension as material, and should note that the claims do not recite "attractor selection" by that name in every instance — they recite "attractor selection control" in the preambles.

Cited references of record

• Patent citations (3): WO2010084775A1 (Panasonic, relay apparatus); JP2011155508A (NTT, virtual network control — the Applicant's own acknowledged Patent Literature 1); US20130039214A1 (Ericsson, OSPF in split‑architecture networks). Earliest priority of cited art: 2009‑01‑26.
• Non‑patent citations (3): the PCT/JP2013/003261 International Search Report (mailed 2013‑08‑27); Leibnitz & Murata, IEEE Network 24(3):14‑18 (May/June 2010); Motoyoshi, Wakamiya & Murata, WONS 2012, pp. 27‑30.
• Later citations of this family: CN104022859A (Beijing Univ. of Posts and Telecom, 2014) and JP2021180389A (NEC, 2021) — indicating continuing NEC interest in this line of work.

Litigation / CAFC docket status — limited confidence

I could not verify any Federal Circuit appeal or district‑court assertion involving US 9,635,675, and I found no 2026 CAFC docket activity for it. Specific caveats, stated honestly:

• I do not have direct authenticated access to USPTO PatentCenter/Patent Trial and Appeal Board or CAFC docket databases; my searches ran against public web indexes (Google Patents, litigation‑document aggregators, dockets.justia, Justia, RPX Insight, PatSnap). A docket could exist that is not surfaced in those indexes.
• Search hits for NEC patent litigation (e.g., ReadyComm LLC v. NEC 2:25‑cv‑304, E.D. Tex.; Deep Green v. NEC; Far North Patents v. NEC 4:19‑cv‑943; e‑Beacon LLC v. NEC 2:24‑cv‑891; NEC v. Anker 2:24‑cv‑720) concern different patents — the asserted numbers in those hits are 9,179,011; 7,714?/(others); the '714, '175, '494, '700 patents; and 11,537,814 respectively. None of the retrieved documents assert 9,635,675.
• The Google Patents record for 9,635,675 shows only routine prosecution‑side events (assignment, notice of allowance/patent grant, and the two maintenance‑fee payments). No adverse legal events, expiration, or reexamination/PGR/IPR certificates appear on that record.
• Consequently: treat "no known litigation or 2026 CAFC docket activity" as a negative search result, not a certified clearance. Verify against PACER and the CAFC docket system before relying on it for anything consequential.

Other number‑deconfliction notes

• RU2606966C2 surfaced in the "similar documents" list with a nearly identical title pattern ("Method of path control, wireless communication system, path control device and not short‑term computer‑readable medium"), but the retrieved Russian text shows it concerns adaptive modulation / FWA backhaul path selection — a different NEC invention, not a family member of 9,635,675. I did not auto‑merge them.
• JP6065011B2 and WO2014038111A1 are genuine family members (same priority JP 2012‑197148, same filing date 2013‑05‑22).

Uncertainty summary: bibliographic data, claims, abstract, and cited art are high confidence (grounded in the authoritative full text and corroborated by mirrors). Litigation/appellate status is low confidence and should be independently verified in PACER and the CAFC docket system.

I'll search for litigation involving US patent 9635675 across patent litigation databases and court sources.

Let me search more targeted litigation databases for this specific patent number.

Let me run a few more targeted searches on litigation-specific databases.

Litigation Search Results for US Patent No. 9,635,675

Bottom line: I found no litigation involving US Patent No. 9,635,675 (the "Motoyoshi" attractor‑selection patent) in the sources I searched. I cannot confirm any district court case, Federal Circuit appeal, ITC investigation, or PTAB proceeding in which this specific patent was asserted or challenged.

What I searched

• General web queries for "9635675 litigation," "9,635,675 patent lawsuit," and NEC + patent infringement + district court complaints.
• Patent-litigation aggregators surfaced in results (Unified Patents litigation portal, RPX/insight.rpxcorp.com, Justia dockets, Docket Alarm, PACERMonitor, Stanford NPE Litigation Database, PatSnap litigation blog, ex parte complaint analyses).
• Targeted queries combining the patent's owner (NEC Corporation) and inventor (Gen Motoyoshi) with litigation terms.

None of the results referenced US 9,635,675 as a patent‑in‑suit. Note that several results mentioning "9,635,675" were unrelated — e.g., a WIPO document (WO1996035675A1), a Czech government contract ID, and unrelated data sets — and I excluded these as non‑responsive, as instructed.

Important caution — do not conflate with other NEC cases

My searches surfaced other, unrelated NEC patent cases, none of which involve US 9,635,675. For example:

• ReadyComm LLC v. NEC Corporation, No. 2:25‑cv‑00304 (E.D. Tex.) — asserted U.S. Patent No. 9,179,011; voluntarily dismissed without prejudice (May 2025).
• NEC Corporation v. Becker Professional Development Corp., No. 3:25‑cv‑00836 / 4:25‑cv‑00334 (N.D. Tex.) — adaptive‑bitrate streaming patents.
• NEC Corporation v. Tone It Up, Inc., No. 1:22‑cv‑00988 (D. Del.) — patents 8,909,809 and 8,752,101.
• VDPP LLC v. NEC, No. 3:24‑cv‑00566 (N.D. Tex.) — 3D‑eyewear patents.

These are cited only to show the pattern of NEC's enforcement activity and to underscore that none of them assert the '675 patent.

About the patent (for identification)

• Patent: US 9,635,675 B2, "Communication device, communication method and non‑transitory computer readable medium."
• Inventor: Gen Motoyoshi; Assignee: NEC Corporation.
• Priority date: 2012‑09‑07 (JP 2012‑197148); Filed: 2013‑05‑22 (PCT/JP2013/003261); Granted: 2017‑04‑25.
• Family members: JP6065011B2, WO2014038111A1, US20150249996A1.
• Legal status: Active; 4th‑ and 8th‑year maintenance fees paid (2020, 2024); anticipated expiration 2033‑05‑22.

Confidence and limitations

I have no high‑confidence basis to assert that litigation exists, and I am explicitly not fabricating a case. Standard caveats:

• This is not a clearance or freedom‑to‑operate opinion. Litigation can exist that is not indexed by public web sources (e.g., recently filed suits not yet docketed in free databases, sealed cases, foreign proceedings, or arbitrations/licensing disputes).
• My searches relied on web‑accessible litigation databases rather than direct PACER, Docket Navigator, or Lex Machina access, which I do not have here.

Recommended verification steps: query PACER directly and dedicated litigation analytics (Lex Machina, Docket Navigator, Unified Patents' litigation portal, and the Stanford NPE database) using the patent number, and check PTAB (IPR/PGR) filings and the USPTO Patent Center's "Litigation" tab / assignment records for any transfers to an assertion entity that might have triggered new suits.

If you can point me to a paid docket source or a specific court, I can help narrow the search further.

I'll verify the structured "no proceedings on file" signal against public sources before writing the report.

PTAB / AIA Trial Analysis — U.S. Patent No. 9,635,675 B2

Patent: U.S. 9,635,675 B2, "Communication device, communication method and non-transitory computer readable medium"
Inventor: Gen Motoyoshi | Assignee: NEC Corporation
Priority: 2012‑09‑07 (JP 2012‑197148) | PCT filed: 2013‑05‑22 | Granted: 2017‑04‑25 | Anticipated expiration: 2033‑05‑22
Source of record: https://patents.google.com/patent/[US9635675](/patent/US9635675)/en

Proceedings overview

Zero AIA trial proceedings are on file. The canonical structured block (USPTO Open Data Portal) returns no Inter Partes Review, Post‑Grant Review, or Covered Business Method review involving this patent, and independent web verification did not surface any proceeding at the PTAB, any Director Review request, or any Federal Circuit appeal arising from an AIA trial. The breakdown is therefore: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials — a clean slate.

Defensive posture: there is no estoppel, no adverse PTAB record, and no narrowing claim amendment to contend with — but also no free gift. All nine claims (1–9, comprising the radio‑interface‑selection species in claims 1, 4, 5 and the domain/cluster‑selection species in claims 2, 3, 6, 7, 8, 9) are untested and presumptively valid. A defendant must build its own invalidity case from scratch. The upside is that every ground is still available and the § 315(b) one‑year bar has never started running against anyone, because the patent has never been served in a complaint that I could locate (see the earlier Litigation summary in this analysis, which likewise found no asserted case — the two findings are mutually consistent).

⚠️ Do not confuse this patent with look‑alike numbers

My searches returned several near‑miss serial numbers that are different patents. Flagging these explicitly so they are not mis‑attributed:

None of these involved NEC, Gen Motoyoshi, or attractor‑selection subject matter.

No proceeding profiles to report

Because the structured list of AIA trials is empty, there are no institution decisions, Final Written Decisions, panels, settlement terminations, or appeals to summarize. I will not manufacture a proceeding number, a panel, or a claim‑level disposition. Any output purporting to give you an FWD citation for U.S. 9,635,675 would be fabricated.

Strategic summary

Claim status. All nine claims are untested and live. There is no cancellation, no certificate of correction narrowing the claims, and no adverse judgment. For a defendant, that cuts both ways: the claims you must beat are the ones as granted, which is generally easier than beating amended claims — but you carry the full burden with no PTAB roadmap to copy. Practically, the two independent‑claim families are:

• Claims 1, 4, 5 — radio‑interface selection: communication environment data = communication quality at a plurality of radio interfaces; fluctuation term = C_k(BER_k − μ_k); select the interface with the largest fluctuation‑equation output. The "shift the average toward zero and multiply by a fixed parameter" limitation is the amendment‑style heart of the claim and is the natural § 103 attack point (it is precisely the artificial‑noise‑to‑measured‑noise substitution described in the specification).
• Claims 2, 3, 6, 7, 8, 9 — domain/cluster selection: communication environment data = inter‑domain control signals; change the domain a node belongs to based on the fluctuation‑equation output and the amount of inter‑domain control signal traffic relative to user traffic.

Estoppel landscape. § 315(e)(2) estoppel is vacuous here — no petitioner, no IPR, no FWD, so nobody is barred from anything. Every prior‑art combination, including all three references the examiner considered, remains fully available to a first‑mover petitioner:

• WO 2010/084775 A1 (Panasonic) — relay apparatus (cited on the face of the patent)
• JP 2011‑155508 A (NTT) — virtual network control method/device, cited as the closest "fluctuation equation" art and expressly identified in the specification as [Patent Literature 1]
• US 2013/0039214 A1 (Ericsson) — OSPF in split‑architecture networks

The examiner also had Leibnitz & Murata, "Attractor selection and perturbation for robust networks in fluctuating environments," IEEE Network 24(3), pp. 14–18 (May/June 2010) and Motoyoshi, Wakamiya & Murata, "Future mobile network management with attractor selection," WONS 2012, pp. 27–30 before the Office as non‑patent literature. Two of these NPL items are authored by the inventor himself and are printed on the face of the patent — which means any § 102(b) attack using the inventor's own 2012 WONS paper runs straight into the one‑year grace period analysis under AIA § 102(b)(1)(A), since the effective filing date is 2013‑05‑22 and the priority date is 2012‑09‑07. Budget for that fight; it is a real limit, not a formality.

Pattern signals. There are none to read. This is not an assertion‑driven patent: no NPE transfer, no defensive aggregator (Unified Patents, RPX, or similar) petition, no multi‑petition serial‑attack pattern, and no PTAB appeal by NEC. Combined with the earlier finding of zero litigation and active maintenance‑fee payments on 2020‑09‑28 and 2024‑10‑16, the picture is an operating company holding a defensive portfolio asset covering a niche control technique (attractor‑selection resource allocation), not a monetization target. That materially lowers the probability that a defendant will be on the receiving end of this patent.

Recommended next steps

1. If you have received a demand or complaint citing U.S. 9,635,675, confirm the assertion from primary sources before doing anything else. There is no litigation I could verify and no PTAB record. Verify via PACER and the USPTO Patent Center "Litigation" tab, and check the assignment record (last recorded assignment: NEC Corporation, 2015‑03‑09, REEL/FRAME 035112/0077) for any transfer to an assertion entity that would explain a newly filed case.

2. If you are planning an IPR, note the timing is wide open. There is no § 315(b) bar (no prior service), no § 325(e) estoppel, and the patent issued 2017‑04‑25, so a PGR is unavailable (§ 321(c), nine months from grant expired 2018‑01‑25) — IPR under § 311 is the only AIA vehicle. Institution decisions are now issued under the bifurcated Director/panel process (USPTO memorandum of 2025‑03‑26) for cases whose POPR deadline falls after that date, which means a discretionary‑denial brief is a separate, earlier‑deadline filing if you are in parallel litigation. Trial must conclude within one year of institution (§ 316(a)(11)); plan your FWD due date accordingly.

3. Consider ex parte reexamination as a lower‑cost first probe. I did not verify ex parte reexamination history here — that is a real gap. Query the USPTO Patent Center "Reexam" tab for this patent before budgeting an IPR.

4. Quote‑and‑link discipline. If you file, cite the granted claims exactly as issued (claims 1–9, with the "fluctuation term" and "largest value" limitations at claim 1 and the "inter‑domain control signals" limitations at claim 2), and cite the patent itself: https://patents.google.com/patent/US9635675/en. If a proceeding is later instituted, the FWD will be public at USPTO PTAB (https://ptacts.uspto.gov/) and any appeal at https://cafc.uscourts.gov / https://www.courtlistener.com — none exists today.

Bottom line: the absence of PTAB activity is itself the signal. This is an unasserted, operating‑company patent covering a narrow biological‑paradigm control technique. Nothing is invalidated, nothing is hardened by PTAB survival, and no estoppel constrains a future petitioner — but there is also no indication anyone has ever needed to test it.

I'll research the assignment record for US 9,635,675, starting with the Google Patents legal events provided (which give me one concrete reel/frame) and then cross-checking against USPTO Assignment Center and other sources.

I found a highly relevant RPX litigation document showing an NEC → "Getner Foundation LLC" patent assignment. Let me examine that and check whether it touches this patent, plus look for the recorded correspondent on reel 035112/0077.

Inventors

Gen Motoyoshi — sole named inventor. The patent's application data (US 14/426,781; US 2015/0249996 A1) lists Motoyoshi only.

• Employer at time of filing: NEC Corporation. The assignment of record (see below) conveys Motoyoshi's rights to NEC Corporation, and Motoyoshi was a NEC researcher at the time — he is the lead author on the NEC/Osaka University attractor-selection papers underlying the patent (G. Motoyoshi, N. Wakamiya, M. Murata, "Future mobile network management with attractor selection," WONS 2012; and the EURASIP JWCN 2012 companion paper). The co-authors (Wakamiya, Murata) are Osaka University academics and are not named inventors, so this is a single-employee-inventor filing.
• Usual-pattern check: No evidence of inventor departure or of a co-inventor exodus. There is no second inventor who could have split the chain. No departure within 12 months is detectable, and none of the classic pre-fire-sale tells appear. I found no indication that Motoyoshi left NEC in connection with this patent.

Original assignee

NEC Corporation (7-1, Shiba 5-chome, Minato-ku, Tokyo 108-8001, Japan) is the assignee named on the issued patent and remains the current assignee per Google Patents.

• Line of business: Operating company — multinational IT/electronics and telecom-network vendor (network infrastructure, 5G, AI, biometrics, system integration). The claimed subject matter (radio interface selection / attractor-selection resource allocation) is squarely within NEC's carrier-network and OpenFlow/future-network R&D activity, which is where the underlying research came from.
• Product embodying the claims: NEC is the employer of the inventor and the R&D sponsor; whether NEC ever shipped a commercial radio-interface-selection product reading on these exact claims is not determinable from the record. What is documented is continued NEC activity in this space: a later NEC family relative, JP2021180389A (priority 2020-05-13, "Communication device, communication system, communication method, and communication program"), filed by NEC and citing this family — evidence NEC kept developing rather than liquidating the technology.
• Current status: Operating. NEC is a publicly listed company (Tokyo Stock Exchange; US ADR). No bankruptcy, dissolution, or acquisition event affects title.

Assignment timeline

The USPTO record for this patent is thin: exactly one recorded assignment — the original inventor-to-NEC conveyance. There is no post-issuance assignment of record, which is itself the key finding: it means NEC (the original assignee) still owns the patent.

• 2015-02-13 (executed / effective) / recorded 2015-03-09 — Reel 035112 / 0077
  • Conveyance: Assignment (Assignment of Assignors' Interest)
  • Assignor: Gen Motoyoshi (individual)
  • Assignee: NEC Corporation (Japan)
  • Correspondent: Not retrievable from the sources I could access. I could not confirm the attorney/firm of record on reel 035112/0077. I will not guess. (Flag: a separate NEC recording — the NEC→Getner Foundation LLC LCD-portfolio assignment recorded at Reel 026312 — shows a submitter named Thomas Watson; I have no evidence he is the correspondent on 035112/0077, and I am not attributing that entry to this patent.)
  • Context: Standard national-stage/employment assignment — the inventor perfects NEC's title as the PCT application (PCT/JP2013/003261) enters the US national stage (US 14/426,781). Not a fire-sale or NPE transfer.

Important cross-reference / contradiction check: My earlier litigation search found no litigation on this patent, and nothing in this assignment chain contradicts that. I did find large NEC→NPE patent transactions in the wild — NEC assigned a large display/LCD patent portfolio to Getner Foundation LLC (an Intellectual Ventures affiliate, 160 Greentree Drive, Suite 101, Dover, DE 19904) via an assignment executed 2011-04-18, and Getner later moved 426 of those assets to Vista Peak Ventures, LLC (recorded 2018-02-28; asserted against AUO/Innolux in E.D. Tex. on 2018-07-10). Those deals concern NEC's LCD/semiconductor patent family (application numbers from the 1990s–2006), not this 2013-filed wireless patent. US 9,635,675 is not on the Schedules I could inspect, and the 2011 execution date precedes this patent's 2013 filing by two years — it could not have been included. I therefore do not place this patent in the NEC→Getner→Vista Peak chain.

Timeline diagram

timeline
    title Ownership of US 9635675
    2012 : Priority application filed in Japan
    2013 : PCT application filed by NEC Corp
    2015 : Motoyoshi assigns rights to NEC Corp
         : Recorded Reel 035112 Frame 0077
    2017 : US patent granted to NEC Corp
    2020 : Fourth year maintenance fee paid
    2024 : Eighth year maintenance fee paid

NPE / troll-pattern signals

1. Shell-entity transfer — Not present. No assignment to any "IP / Patents / Licensing / Holdings / Ventures" LLC appears in the record. The only recorded assignee is NEC Corporation, an operating company.
2. Known asserter in the chain — Not present. The sole assignee of record (reel 035112/0077) is NEC Corporation, which does not match Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Round Rock, or any Unified/RPX high-frequency plaintiff list. Note the near-miss: NEC's separate LCD portfolio did pass to IV affiliate Getner Foundation LLC — but that is a different portfolio and does not include this patent.
3. Repeat correspondent across the chain — Unclear. There is only one link in the chain, so "recurrence" cannot be shown even in principle, and I could not retrieve the correspondent name for reel 035112/0077. No finding either way.
4. Cascading transfers — Not present. Zero post-issuance transfers; nothing cascaded.
5. Pre-litigation transfer — Not present. No litigation exists (per the prior litigation search), so there is no suit to time a transfer against; and no transfer occurred after issuance regardless.
6. Bankruptcy fire-sale — Not present. No bankruptcy of NEC or of any assignee in the chain.
7. Privateering — Not present for this patent. NEC's documented privateering activity (the IV/Getner and Dominion Harbor/Vista Peak display-portfolio monetization) does not include US 9,635,675 on the schedules I could see. Flagging this as a monitored-but-unconfirmed possibility only: NEC demonstrably sells patent portfolios into NPE channels, so a future or unrecorded transfer of this asset cannot be excluded — but there is no recorded or documentary basis to assert it.
8. Defensive aggregator — Not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN.

Verdict

Insufficient data — specifically, the record contains only the original assignment (Motoyoshi → NEC Corporation, executed 2015-02-13, recorded 2015-03-09 at Reel 035112/0077), with no post-issuance transfers. The sole recorded owner is an operating company (NEC), the inventor is a NEC employee, maintenance fees were paid at the 4th and 8th years (2020-09-28 and 2024-10-16), and no litigation, shell-entity transfer, or known-asserter link exists on this patent. There is no evidence of any NPE/troll pattern, but because the chain is a single link I cannot affirmatively characterize it beyond "original assignee retains title."

Caveat I want on the record: I could not load the live Assignment Center entry for Reel 035112/0077, so the correspondent of record for the only assignment is unverified. Everything above is grounded in the Google Patents legal-events data supplied plus the cross-referenced NEC→Getner/RPX documents; where the correspondent is unknown, I have said so rather than guess.

Verify at: USPTO Patent Assignment Search — https://assignmentcenter.uspto.gov/ and https://assignment.uspto.gov/patent/index.html (search "9635675" or application "14/426,781"). Confirm (a) whether any assignment beyond Reel 035112/0077 exists, and (b) the correspondent name on that reel/frame.

I'll verify the citation record and pull details on each cited reference, then assess anticipation.

Let me pull details on the other two patent citations and the non-patent literature.

Prior-Art Analysis — US 9,635,675 B2 (Motoyoshi / NEC)

Bottom line up front

I re-pulled the patent's citation record and verified each cited reference against independent sources (Google Patents' mirror of the USPTO record, the EPO publication server, the JPO published-application PDF, IEEE/ACM for the journal art). Building on the earlier sections, the targeting set is exactly three patent citations and three non-patent citations of record:

§ 102 conclusion: No cited reference anticipates any of claims 1–9. Each of the three patent citations lacks at least the single dispositive limitation that defines the invention — a fluctuation term η′ₖ computed from observed communication‑environment data by subtracting a history average and scaling by a parameter (spec. Eq. 7: η′ₖ = Cₖ(BERₖ − μₖ)). The two most substantive references (P1, P2) actually instantiate the opposite approach the patent attacks (independent random noise), so as § 102 art they are deficient, and as § 103 art they are the references the Applicant had to design around. Details and element mapping below.

⚠️ Verification method caveat: I do not have authenticated USPTO API/PACER access in this session. "USPTO database" results below are drawn from the Google Patents US‑record mirror (https://patents.google.com/patent/US9635675/en), Justia, FreePatentsOnline, and the EPO/JPO servers. The supplied full text remains authoritative where anything conflicts.

Reference P1 — WO2010084775A1 (the closest art)

Full citation: WO 2010/084775 A1, "Relay apparatus, control method, and program" (中継装置、制御方法、およびプログラム), applicant Panasonic Corporation; EP family member EP2293642A1. Priority 2009‑01‑26; published 2010‑07‑29.
Sources: https://patents.google.com/patent/WO2010084775A1/en ; http://data.epo.org/publication-server/rest/v1.2/patents/EP2293642NWA1/document.html

Brief description. A relay device in a multi‑path network that must switch from a primary path to a secondary path on link failure while minimizing maintenance overhead. It discloses (claim 1 of that document): a link status receiving unit receiving link‑status notification messages; a probability generating unit that generates a probability value according to a predetermined probability distribution; a path selection unit that selects one of the paths based on (i) the number of link‑status notification messages and (ii) the probability value; and an "activation [activity] level calculating unit" that computes an activity level indicating suitability of the currently selected primary path, from a ratio of notification counts on non‑primary paths to those on the primary path.

Why the examiner cited it / relevance. This is the record's only reference that squarely discloses an attractor‑selection‑style activity‑plus‑randomness selection mechanism in a communication network: an "activity" metric fed by observed network conditions, combined with a stochastic probability draw, driving discrete path selection. The vocabulary overlap with the '675 specification (activity α, fluctuation, probabilistic selection among alternatives) is strong.

§ 102 analysis — which claims it potentially anticipates: None.

• It discloses an "activity" and a random probability draw, but the probability value comes from "a predetermined probability distribution" — i.e., independent random noise, which is precisely the configuration the '675 patent distinguishes over. It therefore does not teach or disclose "calculating a fluctuation term … by using the communication environment data as the communication environment data occurs," nor the recited mean‑shift + amplitude‑parameter computation (claims 1, 2, 4, 5, 6, 8).
• It does not disclose radio interfaces / base stations / communication quality information (claim 1, claim 4, claim 5), nor domain membership changes driven by inter‑domain control‑signal volume (claims 2, 3, 6, 7, 8, 9). Its subject matter is intra‑network multi‑path routing, not multi‑domain re‑homing.
• Closest claim mapping is only partial: the preamble concept of "attractor selection control using a fluctuation equation" (claims 1, 2, 4, 5, 6, 8) and the general notion of resource allocation via a stochastic selection rule. Partial disclosure is insufficient for anticipation under § 102 — every element, arranged as in the claim, must appear in the single reference.

Reference P2 — JP2011155508A (Applicant‑acknowledged "Patent Literature 1")

Full citation: JP 2011‑155508 A, "Virtual network control method and virtual network control device" (仮想網制御方法および仮想網制御装置), applicants Nippon Telegraph and Telephone Corp. and Osaka University; inventors incl. Takashi Miyamura; filed 2010‑01‑27 (JP 2010‑015936); published 2011‑08‑11. Family/related: JP5305302B2 (granted), EP2530887A1, US20120294617A1.
Sources: http://patentimages.storage.googleapis.com/4a/93/d4/9a10eee02a74b8/JP2011155508A.pdf ; http://data.epo.org/publication-server/rest/v1.2/patents/EP2530887NWA1/document.xml

Brief description. Controls a virtual network topology (over a WDM physical network) by attractor selection, storing a "fluctuation equation" (Langevin equation dx/dt = α·f(x) + η), designing its control parameters (e.g., the regulatory matrix Wᵢⱼ), and applying those parameters to transition the solution between attractors when environmental changes (traffic fluctuation, link/node failure) occur. The EPO counterpart expressly states η is a stochastic function satisfying ⟨η(t)⟩ = 0, ⟨η(t)η(t′)⟩ = Dδ(t−t′), with "a typical example of probability distribution … the normal distribution."

§ 102 analysis — which claims it potentially anticipates: None.

• It discloses the premise of claims 1, 2, 4, 5, 6, 8 — attractor selection control using a fluctuation equation, and an "activity/order parameter α" weighting a deterministic control structure against a fluctuation term. That is why the '675 specification cites it as Patent Literature 1 and expressly states the problem it poses: the fluctuation terms "are independent and use random numbers according to a normal distribution," requiring a designer to pre‑tune a standard deviation to expected variation levels.
• It does not disclose the pivotal claim element: generating the fluctuation term from communication environment data as it occurs, with history‑average subtraction and parameter scaling (claims 1, 2, 4, 5, 6, 8; dependent 3, 7, 9). Its η is explicitly statistically independent of the control terms — the opposite of the '675 η′ₖ.
• No radio‑interface / communication‑quality‑based selection (claims 1, 4, 5) and no inter‑domain control‑signal‑driven domain change (claims 2, 3, 6, 7, 8, 9).
• Because it teaches the independent‑random‑noise variant, it is at most background/§ 103 art — and arguably teaches away from the claimed data‑derived correlated noise, which cuts in the patentee's favor.

Reference P3 — US20130039214A1 (least relevant)

Full citation: US 2013/0039214 A1, "Implementing OSPF in Split‑Architecture Networks," Telefonaktiebolaget L M Ericsson (Publ); inventors Neda Beheshti‑Zavareh, Ying Zhang, Kiran Yedavalli; filed 2011‑08‑11 (Ser. No. 13/208,251); published 2013‑02‑14; granted as US 8,559,314 B2 (2013‑10‑15). Continuation US20140016477A1 → US 8,976,673 B2.
Sources: https://patents.google.com/patent/US20130039214 ; https://www.freepatentsonline.com/y2013/0039214.html

Brief description. Implements OSPF across a split‑architecture (SDN/OpenFlow) network divided into multiple "areas" (expressly, "a section … with separate routing similar to a domain"), each administered by a separate controller. The controller learns the area topology, computes shortest paths between border switches, and exchanges a link‑state database with neighbor controllers to enable inter‑area routing without flooding all internal link costs.

§ 102 analysis — which claims it potentially anticipates: None.

• Its only possible relevance is to the multi‑domain, per‑domain‑control‑device framework of claims 2, 3, 6, 7, 8, 9. It discloses domains/areas and control devices managing nodes — but no attractor selection, no fluctuation equation, and no fluctuation term of any kind, so it cannot anticipate even the preambles of those claims as a whole.
• It has no bearing on claims 1, 4, 5 (radio‑interface selection).
• Most likely cited as general evidence of the multi‑domain network context, or as a § 102(e)-type reference on the "domain/area + controller" vocabulary.

Non‑patent citations

*N1 — Motoyoshi, G., Wakamiya, N., Murata, M., "Future Mobile Network Management With Attractor Selection," 2012 9th Annual Conference on Wireless On‑Demand Network Systems and Services (WONS '12), IEEE, Jan. 2012, pp. 27‑30.*
Note on the record: the Google Patents listing renders this NPL's date as "Jan. 2010" in one field while the specification body states "January 2012" and the conference title is "2012 9th Annual Conference." Per your no‑auto‑correction rule, I flag this internal inconsistency rather than silently resolving it; the WONS 2012 conference was held January 2012, so pp. 27‑30 (Jan. 2012) is the consistent reading.
Relevance: this is the inventor's own paper; the '675 specification cites "section III of Non Patent Literature 1" as the description of typical attractor selection. § 102: it describes attractor selection generally, not the claimed data‑derived fluctuation term; it is background/common‑ownership material, not an anticipating reference.

N2 — Leibnitz, K. & Murata, M., "Attractor selection and perturbation for robust networks in fluctuating environments," IEEE Network 24(3):14‑18 (May/June 2010); DOI 10.1109/MNET.2010.5464222.
Brief description: proposes selecting among networks/services based on each network's sensitivity to inherent fluctuations — a perturbation/robustness framework. § 102: it discusses biologically inspired selection under fluctuation, but discloses no communication‑quality fluctuation‑term computation as claimed; background art only.

N3 — International Search Report for PCT/JP2013/003261, mailed 2013‑08‑27.
This is the search report itself, not prior art. It is the procedural vehicle that produced citations P1–P3; it has no independent § 102 effect.

Consolidated § 102 disposition

Most relevant prior art overall: WO2010084775A1 (Panasonic) is the closest — the only cited reference with a stochastic‑selection‑plus‑activity allocation mechanism in a communication network — followed by JP2011155508A (NTT/Osaka), the Applicant's acknowledged attractor‑selection art. Both were overcome during prosecution because neither derives the fluctuation term from live communication‑environment data.

Statutory framing and caveats

• Anticipation standard: § 102 requires a single reference disclosing every limitation arranged as claimed. On the record retrieved, no such reference exists among the citations; the citations function as § 103 / background art, not anticipatory art. I state this as a negative finding, not a clearance.
• Pre‑AIA vs. AIA: The application's effective filing date for its claims is the JP priority date 2012‑09‑07 (JP 2012‑197148), which is before 16 March 2013, so the pre‑AIA § 102/103 framework presumptively governs (transitional provision, AIA § 3(n)(1)), provided no claim ever had an effective filing date on/after 16 March 2013. I did not verify the prosecution history on that point. If AIA applies instead, the subsection labels change (post‑AIA § 102(a)(1)/(a)(2)) but the substantive outcome here is the same. I flag this as an assumption, not a verified conclusion.
• § 102(b) critical date: For a national‑stage application the U.S. filing date is the international filing date (2013‑05‑22), making the one‑year critical date 2012‑05‑22; P1 (2010), P2 (2011), and N2 (2010) all predate it and are potential statutory‑bar art; N1 (Jan. 2012) and P3 (filed 2011‑08‑11 / published 2013‑02‑14) sit in different categories. Treat the exact categorization as analytical, subject to the pre‑AIA/AIA question above.
• Identifiers not auto‑corrected: WO2010084775A1's EP family member is EP2293642A1; P2's granted family member is JP5305302B2, not to be confused with JP6065011B2 (the granted counterpart of the '675 patent itself). US20130039214A1 granted as US 8,559,314 (not the '675 patent). These are distinct documents and are kept separate here.
• Scope of my search: I looked for prior art beyond the citation record (searching the attractor‑selection / bio‑inspired‑networking literature and NEC/Osaka/NTT patent families) but found nothing indexed in the accessible sources that discloses the specific η′ₖ = Cₖ(BERₖ − μₖ) mechanism with an earlier effective date. That is a negative search result across public web indexes, not a certified completeness guarantee — a full validity study should run a professional database search (e.g., Derwent/STN/PatBase) and confirm the prosecution‑history record.

I'll research the prior art references of record before building the § 103 analysis.

Let me pull details on the two key non-patent references and the Ericsson citation.

Let me verify the family relationship of the NTT reference and pull the Ericsson citation.

Obviousness Analysis — US 9,635,675 B2 under 35 U.S.C. § 103

Building on the previously generated sections. The summary section already established the claim set, the Eq. (7) core (η′ₖ = Cₖ(BERₖ − μₖ)), the cited-art list, and two flagged tensions I reuse here: (i) the granted claims do recite multiplying by a "given parameter … to adjust an amplitude," which sits uneasily with the "no amplitude tuning needed" benefit; and (ii) claim 6 contains an apparent drafting error ("generating communication environment data generating device"). Both matter below. I do not repeat that material. This section adds the § 103 mapping and motivation analysis.

Scope and honesty note. I do not have full-text authenticated copies of four of the five references; my disclosure characterizations rely on (a) the supplied authoritative patent text, (b) the applicants' own characterizations of the references in the specification, and (c) retrieved abstracts/selected passages with the URLs cited below. The actual teachings of the references' full bodies must be confirmed against paper/PDF copies before this analysis is relied on. This is an analyst's hypothetical, not a legal opinion.

1. Legal framework applied

• Graham v. John Deere Co. v. Laitram, 383 U.S. 1 (1966): scope and content of the prior art; differences between the prior art and the claims; level of ordinary skill; secondary considerations.
• KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007): a claimed combination is obvious where the elements "work no differently than they do separately"; a "known technique" used to improve a similar device is obvious; "if a person of ordinary skill can implement a predictable variation, § 103 likely bars its patentability"; and a finite set of identified, predictable solutions yields an "obvious to try" rejection.
• In re Fulton / Medichem v. Rolabo: mere preference for a different embodiment is not teaching away; teaching away requires criticism, discrediting, or discouragement of the claimed solution.
• Applicant-admitted prior art: a specification's characterization of "existing"/"related" systems is usable as prior art and as evidence of the knowledge of a POSITA. This is central here — see §3.

Level of ordinary skill (proposed). A POSITA would hold an M.S. or Ph.D. in EE/CS or equivalent, or a B.S. plus 3–5 years, in network control; would be familiar with stochastic dynamical systems and the attractor-selection paradigm (the fluctuation equation dm/dt = α·f(m) + η and its biological basis); and would possess routine statistical signal-processing competence (time-series statistics, mean-centering, variance/amplitude scaling). This is the level against which the references are read.

2. Element-by-element mapping of the independent claims

Observation that drives everything below: every element except C/E is disclosed or admitted prior art. The § 103 question therefore collapses to: would it have been obvious to replace the injected artificial noise term with a fluctuation term derived from measured communication-environment data?

3. The prior art, as actually characterized

NPL 1 — Motoyoshi, Wakamiya & Murata, WONS 2012 (also cited in the '675 spec as Non Patent Literature 1)

Source: https://ieeexplore.ieee.org/document/[6152230](/patent/6152230) and https://researchmap.jp/Masayuki-Murata/published_papers/24734050

This is the single most damaging reference, because the '675 specification reproduces its equations essentially verbatim:

• WONS Eq. (2): dmᵢ/dt = s(α)/(1+max(mᵢ)² − mᵢ²) − d(α)·mᵢ + ηᵢ, with Eq. (3) s(α) = α(βα^γ + 1/√2), d(α) = α.
• '675 Eq. (8): dmₖ/dt = α(βα^γ + 1/√2)/(1 + max(m₁,…,m_M)² − mₖ²) − α·mₖ + η′ₖ.

Those are the same equation. WONS discloses "mᵢ indicates selection fitness of each radio interface of a mobile node," computed from real-time SNR and available bandwidth with packet-loss statistics; and the clustering variant where "sⱼ indicates selection probability of a cluster" computed as the user-data/signaling-traffic ratio. The only difference is that WONS's ηᵢ "is white Gaussian noise" — exactly the deficiency the '675 spec concedes.

Notably, the applicant admits this is prior art: the '675 spec says "FIG. 2 illustrates an example of a radio interface selection control system configuration based on related attractor selection control"; FIG. 3 is an "existing communication system"; and the Background states that the fluctuation terms "given in the literatures are independent and use random numbers according to a normal distribution." That is a judicial admission of the entire architecture of claims 1, 2, 4, 5, 6, 8 minus the noise-term derivation.

NPL 2 — Leibnitz & Murata, IEEE Network 24(3):14–18 (May/June 2010)

Source: https://ieeexplore.ieee.org/abstract/document/[5464222](/patent/5464222) ; https://dl.acm.org/doi/abs/10.1109/MNET.2010.5464222

This reference supplies the missing element. Its abstract:

"we propose a new framework for selecting among different networks and services based on the robustness of each network's performance metrics. The selection is only based on observations of the system's responsiveness to inherent fluctuations. … Due to the explicit utilization of the inherent fluctuations in the system, the proposed selection scheme can operate smoothly to select the most suitable and robust network."

That is precisely "generating fluctuation information by using the communication environment data as the data occurs" and "a fluctuation term reflecting the communication environment data." It was published more than two years before the 2012‑09‑07 priority date, by authors who are not the '675 inventor (so no § 102(b)(1)(A) self-disclosure issue), and it shares a co-author lineage (Murata) with the WONS work — same research program, same field.

Patent Literature 1 — JP 2011‑155508 A (NTT + Osaka University)

Sources: http://patentimages.storage.googleapis.com/4a/93/d4/9a10eee02a74b8/JP2011155508A.pdf ; https://jglobal.jst.go.jp/en/detail?JGLOBAL_ID=201103000910783185 ; family: https://patents.google.com/patent/JP5305302B2/en ; https://patents.google.com/patent/EP2530887A1 (family: priority JP 2010‑0015934/‑015936/‑015939, filed 2010‑01‑27; WO2011JP51643; CN102726011A; US 2012/0294617 A1) — the family relationship is confirmed by the shared priority numbers listed on Espacenet.

Discloses the attractor-selection fluctuation equation framework for allocating communication resources (wavelength paths / virtual network topology), with the store-the-equation / design-the-parameters / apply-to-control steps (its claim 1), the order parameter α, and — critically — the defining constraints on the noise term:

<η(t)> = 0, <η(t)η(t′)> = Dδ(t−t′), <g(x₀)η(t)> = 0 … "A typical example of probability distribution of the function η satisfying these conditions is the normal distribution."

and discloses applying the fluctuation equation "periodically or every time the environmental fluctuation does occur" (US 2012/0294617 A1, ¶[0047]). The zero-mean constraint <η> = 0 is the very thing the claim's "shifting an average value of fluctuations toward zero" is designed to satisfy.

Other cited art

• WO 2010/084775 A1 (Panasonic, pub. 2010‑07‑29): relay path selection using "a probability generating unit which generates a probability value according to a predetermined probability distribution," a "path selection unit which selects one of the plurality of paths based on the number of the link status notification message and the probability value," and an "activation level … based on a ratio" of received messages — https://patents.google.com/patent/WO2010084775A1/en. Useful as evidence that (i) stochastic selection driven by an observed network-state ratio, and (ii) a tunable amplitude/probability parameter, were known in this field.
• US 2013/0039214 A1 (Ericsson, "Implementing OSPF in Split-Architecture Networks," pub. 2013‑02‑14). ⚠️ I was unable to retrieve the substance of this document within my search budget; I therefore do not assign it any specific teaching in the combinations below. It is relevant only on its face as evidence that area/domain structure and control-plane signaling in split-architecture networks were known.

4. The § 103 combinations

Combination I — WONS 2012 (NPL 1) in view of Leibnitz & Murata (NPL 2) → claims 1, 4, 5

(and, with NPL 1's clustering subsection, claims 2, 6, 8 and dependent claims 3, 7, 9)

Difference from the claims: the sole substantive difference is element C/E — NPL 1 uses white Gaussian ηᵢ; the claim requires the fluctuation term to be derived from the communication-environment data (mean-shifted, amplitude-scaled).

Motivation to combine, articulated:

1. Same field, same problem. Both references address adaptive/robust network control under fluctuating environments; NPL 2 is expressly about "selecting among different networks and services," i.e., exactly NPL 1's multi-interface/cluster selection problem. KSR ("familiar elements … according to known methods").
2. NPL 2 provides the express suggestion. It does not merely permit the substitution; it advocates it — selection "only based on observations of the system's responsiveness to inherent fluctuations," with the stated advantage that the scheme "can operate smoothly." A POSITA looking to reduce the manual standard-deviation tuning that NPL 1 and Patent Lit. 1 both require would be led directly to NPL 2's approach. KSR ("a need or problem known in the field … can provide a reason").
3. The residual arithmetic is routine. Once one uses measured data as the fluctuation source, satisfying the accepted design constraint <η> = 0 (expressly recited in Patent Lit. 1) is done by mean-centering — subtracting the historical average from the instantaneous observation. Multiplying by a constant (Cₖ) to set amplitude is nothing more than the ordinary scaling of a noise term's standard deviation, which every reference already does (NPL 1's white Gaussian has a variance parameter). The '675 spec itself describes these two steps in purely conventional statistical terms ("shift the average of fluctuations toward 0," "adjust amplitude"). KSR: predictable variation.
4. Obvious to try. The art identified only two candidate noise sources — injected artificial noise or the system's own observed fluctuation — and NPL 2 identifies the latter as the improvement. KSR ("finite number of identified, predictable solutions").
5. Reasonable expectation of success. NPL 1 had already demonstrated the control loop works on live SNR/BW/packet-loss statistics; swapping the noise source leaves the demonstrated mechanism intact.

Result: claim 1 (and its device/program counterparts 4 and 5) rendered obvious. Claims 2/6/8 and 3/7/9 follow from NPL 1's own clustering disclosure, with element G being almost the verbatim teaching of NPL 1 ("the ratio of user data traffic volume against signaling traffic volume … passing through the network device").

Combination II — Patent Literature 1 + NPL 2 + WO 2010/084775 (+ NPL 1 as admitted prior art) → all claims

This is the belt-and-suspenders combination that survives even if NPL 1 were removed as § 102 antecedent art (see § 5, issue 4).

• Patent Lit. 1 supplies (a) the attractor-selection fluctuation-equation method for allocating communication resources, (b) the store/design/control step structure, (c) the <η> = 0 design constraint, and (d) the "every time the environmental fluctuation does occur" timing that maps onto "as the data occurs."
• NPL 2 supplies the substitution to inherent/observed fluctuations.
• WO 2010/084775 supplies stochastic selection driven by an observed network-state ratio and a scaling/"probability" parameter, corroborating "multiplying a given parameter."
• NPL 1 supplies the radio-interface and cluster/domain applications, and independently remains available as applicant-admitted prior art because the '675 specification itself identifies it as the state of the art and reproduces its equations.

Motivation is the same as Combination I, with the added rationale that all four references are directed to the shared problem of network control under environmental fluctuation and are therefore "reasonably pertinent to the particular problem with which the inventor was concerned."

Combination III — any of the above + US 2013/0039214 A1 (Ericsson)

Relevant only to claim 2/6/8's "communication network made up of a plurality of domains" and inter-domain control signaling. ⚠️ Because I could not retrieve its content, I flag this as unverified and do not rely on it.

5. Anticipated counterarguments and rebuttals

1. "Teaching away — the field consistently used artificial Gaussian noise."
NPL 1 uses white Gaussian ηᵢ; Patent Lit. 1 teaches η is "a random number according to a normal distribution." Rebuttal: disclosure of a preferred embodiment is not teaching away. No reference "criticize[s], discredit[s], or otherwise discourage[s]" data-derived fluctuation — and NPL 2 affirmatively teaches it. Under In re Fulton and Medichem, this argument fails.

2. "Unexpected result — the invention eliminates noise-amplitude tuning."
This is the strongest non-obviousness story the applicant can tell, and the specification tells it in its Advantageous Effects section. But it collides with the claim language: claim 1 itself requires "adjusting an amplitude of the shifted fluctuation term … by multiplying a given parameter into the shifted fluctuation term." So construed, the claims retain an amplitude parameter — the very thing the asserted advantage says is eliminated — and the asserted benefit is not co-extensive with the claim scope, defeating the required nexus for any secondary-consideration argument. (Construed the other way, to avoid the tuning parameter, the claims read directly onto the admitted prior art.) This is the same tension flagged in the summary section; it is material to validity and, secondarily, to § 112.

3. "The claim requires the fluctuation term as it occurs / in real time."
Satisfied by NPL 1's "instantaneous live traffic volume" and "real-time SNR," and by Patent Lit. 1's "every time the environmental fluctuation does occur."

4. "NPL 1 is the inventor's own disclosure and is excepted under AIA § 102(b)(1)(A)."
Genuine nuance: the '675 application's effective filing date is 2012‑09‑07 (post-AIA), and the WONS paper appeared January 2012 — within one year. But the paper is a joint work by Motoyoshi plus two non-inventors (Wakamiya, Murata), and § 102(b)(1)(A) excepts only disclosures by "the inventor or joint inventor," or by "another who obtained the subject matter disclosed directly or indirectly from the inventor or a joint inventor." Whether co-authors who are not inventors qualify raises a fact question. This issue is ultimately non-dispositive, for two reasons: (i) the specification's own Background/"related techniques" discussion is an admission that the FIG. 2/FIG. 3 architecture and the fluctuation-equation approach were existing art, and admissions are not erased by § 102(b)(1)(A); and (ii) NPL 2 (Leibnitz & Murata, 2010) is independently unquestionable prior art and supplies the key teaching. Patent Lit. 1 (published 2011‑08‑11, priority 2010‑01‑27) is also unquestionable art.

5. "Secondary considerations."
Nothing in the record (as surveyed in the earlier section) evidences commercial success, long-felt but unsolved need, failure of others, licensing, or copying tied to the claims. To the contrary, the existence of the inventor's own WONS 2012 paper undercuts any long-felt-need narrative — the inventors were already at the claimed solution before the priority date.

6. Dependent claims (3, 7, 9) are the weakest.
Claim 3's "control formula … represented by using an amount of control signal traffic with respect to an amount of user traffic passing through the communication node" tracks NPL 1's "ratio of user data traffic volume against signaling traffic volume" almost word for word. If the independent claims fall, these fall a fortiori.

6. Bottom line and confidence

The decisive structural point: the '675 specification concedes that everything except the derivation of the noise term from measured environment data is admitted prior art, and NPL 2 — published 27 months before the priority date by different authors in the same research field — expressly teaches selecting based on "the inherent fluctuations in the system." That gives a clean KSR combination: a known technique (data-derived fluctuation) applied to a known system (attractor-selection interface/cluster control) to obtain the predictable result of avoiding manual noise-variance tuning.

Contradiction check against the earlier sections: none found. The two flags from the summary (the amplitude-parameter tension; the claim 6 drafting error) are consistent with — and in the case of the amplitude parameter, central to — this analysis. The earlier "no known litigation" finding is unaffected and not relied upon here.

Caveat repeated: this rests on abstracts and selected passages, not full authenticated copies of NPL 1, NPL 2, Patent Lit. 1, and WO 2010/084775, and on no content at all from US 2013/0039214 A1. Verify against full texts (and, for the § 102(b)(1)(A) question, against the inventors' and co-authors' contributions) before relying on this for any consequential purpose.