# US 2024/0391194 A1 · Machine learning for splice improvement

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2024/0391194 A1
- Document ID: US20240391194A1
- Status: Published application
- Kind: A1
- Application number: 18/792,217
- Filed: 2024-08-01
- Published: 2024-11-28
- Earliest claimed: 2019-07-12
- Assignees: Bridgestone Americas Tire Operations, LLC
- Applicants: Bridgestone Americas Tire Operations, LLC
- Inventors: Kevin R. Knapp; Anirban K. Bhattacharjee
- Web page: https://patentgrep.com/patent/US20240391194A1

## CPC

- B29D 30/005: {General arrangement or lay-out of plants for the processing of tyres or parts thereof (vulcanization presses B29D30/0601; round cores or cylindrical drums arranged for a single sequence of tyre building operations B29D30/10, B29D30/20)}
- B29D 30/0061: {Accessories, details or auxiliary operations not otherwise provided for}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- B29D 2030/0055: {Optimization of the cycle times of the tyre manufacturing process, e.g. adaptation of the tyre building process to the vulcanization process}
- B29D 2030/0066: {Tyre quality control during manufacturing}
- G06N 3/045: Combinations of networks
- G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections
- G06N 3/084: Backpropagation, e.g. using gradient descent
- B29D: PRODUCING PARTICULAR ARTICLES FROM PLASTICS OR FROM SUBSTANCES IN A PLASTIC STATE (making granules B29B9/00; making preforms B29B11/00)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system to manufacture a tire includes one or more processors and one or more memories storing instructions executable by the one or more processors and causing the one or more processors to receive, from one or more sensors of one or more pieces of tire manufacturing equipment, one or more values corresponding to manufacturing of a tire, generate a matrix based on the one or more values, predict, via input of the matrix into a machine learning model, a value for a splice tolerance metric for the tire, determine, based on the value for the splice tolerance metric, a first parameter of a first piece of the one or more pieces of tire manufacturing equipment to adjust, and provide a command to adjust the first parameter of the first piece of equipment responsive to determining the value of the splice tolerance metric.

## Claims

### Claim 1

A system to manufacture a tire, the system comprising:
one or more processors and one or more memories, the one or more memories configured to store instructions that are executable by the one or more processors and cause the one or more processors to:
receive, from one or more sensors of one or more pieces of tire manufacturing equipment, one or more values corresponding to manufacturing of a tire;
generate a matrix based on the one or more values;
predict, via input of the matrix into a machine learning model, a value for a splice tolerance metric for the tire;
determine, based on the value for the splice tolerance metric, a first parameter of a first piece of the one or more pieces of tire manufacturing equipment to adjust; and
provide a command to adjust the first parameter of the first piece of equipment responsive to determining the value of the splice tolerance metric.

### Claim 2 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to determine one or more metrics based on the one or more values, wherein generating the matrix is further based on the one or more metrics.

### Claim 3 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to construct a plurality of machine learning models for each of the one or more pieces of tire manufacturing equipment paired with each of a plurality of splices of the tire, wherein the machine learning model is one of the plurality of machine learning models.

### Claim 4 (depends on 1)

The system of claim 1, wherein the machine learning model is a first machine learning model of a plurality of machine learning models, the one or more processors further caused to:
construct the first machine learning model of the plurality of machine learning models for a first pressure roller component of the one or more pieces of tire manufacturing equipment and a first splice of the tire;
construct a second machine learning model of the plurality of machine learning models for the first pressure roller component and a second splice of the tire;
construct a third machine learning model of the plurality of machine learning models for the first pressure roller component and a third splice of the tire;
construct a fourth machine learning model of the plurality of machine learning models for a second pressure roller component and the first splice of the tire;
construct a fifth machine learning model of the plurality of machine learning models for the second pressure roller component and the second splice of the tire;
construct a sixth machine learning model of the plurality of machine learning models for the second pressure roller component and the third splice of the tire;
construct a seventh machine learning model of the plurality of machine learning models for a third pressure roller component and the first splice of the tire;
construct an eighth machine learning model of the plurality of machine learning models for the third pressure roller component and the second splice of the tire; and
construct a ninth machine learning model of the plurality of machine learning models for the third pressure roller component and the third splice of the tire.

### Claim 5 (depends on 1)

The system of claim 1, wherein the first parameter is one of a pressure, an orientation, an alignment, or a position value.

### Claim 6 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to train the machine learning model based on a degree to which a previous splice deviated from a predetermined splice point, a moving average of a splice point for a predetermined number of periods, or a slope of a predetermined number of moving averages of the splice point.

### Claim 7 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to:
determine, based on a plurality of machine learning models, that the value for the splice tolerance metric predicts that a first splice of a plurality of splices of the tire exceeds a tolerance window;
select, based on a policy, the first piece of equipment corresponding to the first splice of the plurality of splices of the tire; and
command the first piece of equipment selected based on the policy to adjust a characteristic of the first splice to bring the first splice within the tolerance window.

### Claim 8 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to:
receive, from the one or more pieces of tire manufacturing equipment, a time series of the one or more values;
determine, based on the time series of the one or more values, a moving average over a plurality of predetermined periods, wherein the plurality of predetermined periods comprise at least two different predetermined periods;
determine a slope of the moving average over the plurality of predetermined periods; and
generate the matrix based on the slope of the moving average over the plurality of predetermined periods.

### Claim 9 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to:
identify a confidence score associated with the value for the splice tolerance metric predicted via the machine learning model;
compare the confidence score with a threshold; and
determine to adjust the first parameter of the first piece of equipment responsive to the confidence score satisfying the threshold.

### Claim 10 (depends on 1)

The system of claim 1, wherein the one or more processors are further caused to:
predict, based on an output of the machine learning model, that a splice for the tire does not satisfy a tolerance window, wherein the first parameter is adjusted based on the prediction;
receive, subsequent to the adjustment of the first parameter, one or more updated values corresponding to the manufacture of the tire by the one or more pieces of tire manufacturing equipment;
generate an updated matrix based on the one or more updated values received subsequent to the adjustment of the first parameter;
predict, via input of the updated matrix into the machine learning model, whether the splice for the tire satisfies the tolerance window; and
when the prediction indicates that the splice for the tire does not satisfy the tolerance window subsequent to the adjustment, adjust at least one of a second parameter of the first piece of equipment or a parameter of a second piece of equipment.

### Claim 11

A method of manufacturing a tire, the method comprising:
receiving, by one or more processors from one or more sensors of one or more pieces of tire manufacturing equipment, one or more values corresponding to manufacturing of a tire;
generating, by the one or more processors, a matrix based on the one or more values;
predicting, by the one or more processors, via input of the matrix into a machine learning model, a value for a splice tolerance metric for the tire;
determining, by the one or more processors based on the value for the splice tolerance metric, a first parameter of a first piece of the one or more pieces of tire manufacturing equipment to adjust; and
providing, by the one or more processors, a command to adjust the first parameter of the first piece of equipment responsive to determining the value of the splice tolerance metric.

### Claim 12 (depends on 11)

The method of claim 11, further comprising determining, by the one or more processors, one or more metrics based on the one or more values, wherein the matrix is further generated based on the one or more metrics.

### Claim 13 (depends on 11)

The method of claim 11, further comprising:
constructing a plurality of machine learning models for each of the one or more pieces of tire manufacturing equipment paired with each of a plurality of splices of the tire, wherein the machine learning model corresponds to one of the plurality of machine learning models.

### Claim 14 (depends on 11)

The method of claim 11, wherein the machine learning model is a first machine learning model of a plurality of machine learning models, the method further comprising:
constructing the first machine learning model of the plurality of machine learning models for a first pressure roller component of the one or more pieces of tire manufacturing equipment and a first splice of the tire;
constructing a second machine learning model of the plurality of machine learning models for the first pressure roller component and a second splice of the tire;
constructing a third machine learning model of the plurality of machine learning models for the first pressure roller component and a third splice of the tire;
constructing a fourth machine learning model of the plurality of machine learning models for a second pressure roller component and the first splice of the tire;
constructing a fifth machine learning model of the plurality of machine learning models for the second pressure roller component and the second splice of the tire;
constructing a sixth machine learning model of the plurality of machine learning models for the second pressure roller component and the third splice of the tire;
constructing a seventh machine learning model of the plurality of machine learning models for a third pressure roller component and the first splice of the tire;
constructing an eighth machine learning model of the plurality of machine learning models for the third pressure roller component and the second splice of the tire; and
constructing a ninth machine learning model of the plurality of machine learning models for the third pressure roller component and the third splice of the tire.

### Claim 15 (depends on 11)

The method of claim 11, wherein the first parameter is one of a pressure, an orientation, an alignment, or a position value.

### Claim 16 (depends on 11)

The method of claim 11, further comprising:
constructing the machine learning model based on a degree to which a previous splice deviated from a predetermined splice point, a moving average of a splice point for a predetermined number of periods, or a slope of a predetermined number of moving averages of the splice point.

### Claim 17 (depends on 11)

The method of claim 11, further comprising:
determining, based on a plurality of machine learning models, that the value for the splice tolerance metric predicts that a first splice of a plurality of splices of the tire exceeds a tolerance window;
selecting, based on a policy, at least one piece of equipment corresponding to the first splice of the plurality of splices of the tire; and
commanding the at least one piece of equipment selected based on the policy to adjust a characteristic of the first splice to bring the first splice within the tolerance window.

### Claim 18 (depends on 11)

The method of claim 11, further comprising:
receiving, from the one or more pieces of tire manufacturing equipment, a time series of the one or more values;
determining, based on the time series of the one or more values, a moving average over a plurality of predetermined periods, wherein the plurality of predetermined periods comprise at least two different predetermined periods;
determining a slope of the moving average over the plurality of predetermined periods; and
generating the matrix based on the slope of the moving average over the plurality of predetermined periods.

### Claim 19 (depends on 11)

The method of claim 11, further comprising:
identifying a confidence score associated with the value for the splice tolerance metric predicted via the machine learning model;
comparing the confidence score with a threshold; and
determining to adjust the first parameter of the first piece of equipment responsive to the confidence score satisfying the threshold.

### Claim 20 (depends on 11)

The method of claim 11, further comprising:
predicting, based on an output of the machine learning model, that a splice for the tire does not satisfy a tolerance window, wherein the first parameter is adjusted based on the prediction;
receiving, subsequent to the adjustment of the first parameter, one or more updated values corresponding to the manufacture of the tire by the one or more pieces of tire manufacturing equipment;
generating an updated matrix based on the one or more updated values received subsequent to the adjustment of the first parameter;
predicting, via input of the updated matrix into the machine learning model, whether the splice for the tire satisfies the tolerance window; and
when the prediction indicates that the splice for the tire does not satisfy the tolerance window subsequent to the adjustment, adjusting at least one of a second parameter of the first piece of equipment or a parameter of a second piece of equipment.
