patent.fyi
Applications  /  US 2026/0354431
United States Patent Application Publication

US 2026/0354431 A1

Published
Nov 12, 2026
Filed
May 27, 2026
Earliest date on the face
Dec 31, 2020
Classification
A23L 33/15
First published here
Nov 18, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Manual worksheet and weighing

What ifWhat if an operator prepares each package with printed worksheets and a mechanical balance?
Claim language“a processor”

The operator adds the printed benefit values for the subject’s flagged wellness categories and sorts ingredient cards by the resulting totals. The operator weighs the listed doses in that order, passing over any dose that exceeds the remaining package mass.

1.2Recipes keyed to wellness profiles

What ifWhat if the system retrieves complete recipes by wellness profile without storing ingredient benefit measures?
Claim language“a benefit measure for each of a plurality of wellness categories”

Each record pairs a bit mask of flagged wellness categories with a recipe containing ingredient codes and gram amounts. The system matches the subject’s mask to a record and sends its recipe to the weighing station after checking the total mass.

1.2Subject selected wellness categories

What ifWhat if the subject selects wellness categories directly without taking a scored assessment?
Claim language“wellness categories for which the subject has a suboptimal score”

The intake form lets the subject select up to three categories from a fixed menu. The system uses those selections to retrieve ingredient benefit values for the package calculation.

1.2Direct calculation of ingredient amounts

What ifWhat if a constrained calculation sets ingredient amounts directly without grouping material into bins?
Claim language“bins comprising a different ingredient or combination of ingredients”

The calculation assigns one mass variable to each ingredient and maximizes the total benefit across the selected wellness categories using a benefit coefficient per gram. It caps each variable at its ingredient limit and keeps the sum of all masses within the package limit.

1.2Proportions scaled to package mass

What ifWhat if each candidate specifies a dimensionless mass share and the system calculates dispensing amounts only after selection?
Claim language“the target amount comprises an ingredient dose”

The system adds the shares of the selected candidates and divides each share by that sum. It multiplies each normalized share by the package mass and checks the resulting gram amounts against ingredient limits before dispensing.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Fertilizer blends for greenhouse beds

A controller stores the contribution of each granular fertilizer to the nutrient targets for a greenhouse bed and imports the bed's soil test results. It converts nutrient shortfalls to weights, scores each hopper by its fertilizer's weighted nutrient contributions, and calculates a target mass from the fertilizer composition and bed area. It visits hoppers in descending score order and adds each target mass to the batch recipe when it fits within the remaining bag capacity.

Component substitution

Sparse matrix for ingredient benefit data

A compressed sparse row matrix replaces the ingredient benefit database in memory, with one row per ingredient and one column per wellness category. A separate bin table stores ingredient row indices and ingredient masses at each bin's target dose. The processor multiplies the matrix by a vector of subject category weights, combines the resulting ingredient scores by bin, and selects target doses in descending bin score order within the package mass limit.

Parameter or scale variation

Dose selection on a fixed mass grid

The controller stores validated target amounts from 100 to 1,000 milligrams in 25 milligram increments and a package limit of 4,000 milligrams. It represents all amounts as integer counts of 25 milligram units and initializes the remaining capacity to 160 units. It scans the bins in descending utility order and subtracts a bin's target count only when that count fits within the remaining capacity.

Combination with existing systems

Recipe exchange with a batch historian

The formulation controller sends each ranked recipe to a manufacturing execution system with a package identifier, ingredient bin identifiers, and target masses. The manufacturing execution system resolves each bin to its active lot and records measured dispensed masses from the packaging line in a batch historian under the package identifier. A reconciliation service joins those records to the recipe and flags any dose outside its configured mass tolerance.

Automated or data-driven variant

Questionnaire scoring with a calibrated model

A tablet collects responses to a fixed set of wellness questions and passes numeric responses to one logistic regression model per wellness category. Each model uses coefficients fitted to labeled questionnaire records and returns a probability of a low category score. The controller uses probabilities above configured category thresholds as weights for ingredient benefit measures, ranks the bins by the resulting utility scores, and selects target doses that fit the package mass limit.

Inventory sensing before package assignment

Load cells beneath ingredient hoppers report gross mass to the formulation controller, which subtracts each hopper's tare and minimum operating reserve to calculate usable stock. Before scanning the utility-ranked list for a package, the controller subtracts quantities reserved for queued packages and excludes bins with less usable stock than their target dose. It reserves selected doses in a single database transaction and reconciles those reservations against measured consumption when the packaging line reports completion.

Specific version of a part

Embedded processor for package calculation

A 32-bit microcontroller calculates bin utility scores and builds the package ingredient list. It sums target masses as integer milligrams and compares the running total with the configured package limit.

Flash stored benefit matrix

Nonvolatile flash memory stores a two-dimensional table with one row per ingredient and one column per wellness category. Each cell holds a 16-bit integer benefit value that the processor reads using the ingredient and category indices.

Utility weighted by dose mass

For each bin, the processor sums the benefit measures for the subject’s suboptimal categories separately for each ingredient. It multiplies each sum by that ingredient’s target dose in grams and adds the products to obtain the bin utility score.

Ranked filling with mass checks

The processor sorts bins by descending utility score and uses the bin identifier to resolve ties. It traverses that order and adds a bin’s full target amount whenever it fits within the remaining package mass.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 2006/0099310 A1
Published
May 11, 2006
Filed
Nov 10, 2004
Class
A23L 3/36
Document BUS 2012/0144160 A1
Published
Jun 7, 2012
Filed
Dec 7, 2010
Class
G06F 15/76
Claim 1
Document A
Document B
1.1(a) a computer comprising: (i) a processor; and
The processor/controller can he a programmable logic device (PLD), microcontroller, or computer, depending on the nature of the duties required by the processor/controller.
A · US 2006/0099310 A1 · ¶ 25
A multi-cycle programmable processor method implemented in a computer having microprocessor functional units connected to a controller selectively controlling the functional units, thereby controlling datapaths associated with the functional units, the method comprising the step of implementing a finite state machine within the controller, the finite state machine having a plurality of states including an instruction fetch state, an instruction decode state, and an instructio
B · US 2012/0144160 A1 · claim 6
1.2(a) a computer comprising: (ii) a memory, coupled to the processor, the memory storing a module comprising: (1) a database comprising, for each of a plurality of ingredients, a benefit measure for each of a plurality of wellness categories; (2) data from a subject including one or a plurality of wellness categories for which the subject has a suboptimal score; (3) an algorithm which, when executed, calculates for each of a plurality of bins comprising a different ingredient or combination of ingredients, a utility score, wherein the bin utility score is a function of the benefit measures for the suboptimal wellness categories of each ingredient in the bin, and a target amount of material in the bin, wherein the target amount comprises an ingredient dose of one or more of the ingredients; (4) an algorithm which, when executed, determines one or more ingredient sets, each ingredient set comprising material from each of a plurality of the bins, wherein the sum of the target amounts of material in the set is no greater than a total package amount, and wherein the materials in the set are selected from the bins in rank order beginning with the bin having the highest utility score; and (5) computer executable instructions for implementing the algorithms.
The processor/controller can access the memory/database component of the system to run one or more appropriate programs or algorithms to translate the personal data provided into a recipe for a personalized nutritional supplement and directions for compounding a personalized nutritional supplement. The personal data provided and the programs or algorithms necessary to process that information into a personalized nutritional supplement are stored in the memory/database.
A · US 2006/0099310 A1 · ¶ 25

An empty cell means only that no passage was reproduced there.

Statements in these documents
In a preferred embodiment, the disclosed invention relates to devices and methods for automatically preparing personalized nutritional supplements in a store or other retail center.
A · US 2006/0099310 A1 · ¶ 1
The disclosed invention improves upon this system by providing an automated device for formulating a personalized nutritional supplement.
A · US 2006/0099310 A1 · ¶ 8
The time period of a clock can be significantly reduced in this implementation compared to single cycle implementation, which, in turn, will speed up the processing time.
B · US 2012/0144160 A1 · ¶ 6
As the mix of instructions varies considerably based upon the type of processing application being run, it would be desirable to de-couple the performance improvement from the mix of instructions in the program being executed.
B · US 2012/0144160 A1 · ¶ 9
Cite aspatent.fyi, US 2026/0354431 A1, claim 1, first published 2026-11-18T22:28:45-05:00, https://patent.fyi/US20260354431/, SHA-256 fe47d05d89268b41c369fe8fd69a8ad994618936f6c90e07518b7cb1a9a2a29d
First published
Permanent addresshttps://patent.fyi/US20260354431/
Content fingerprint (SHA-256)fe47d05d89268b41c369fe8fd69a8ad994618936f6c90e07518b7cb1a9a2a29d
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi