Granted Patent B2

US 12,505,359 · App. 17/806,021 · Granted Dec 23, 2025

Inventors: Hendrik Schoeniger (Dreieich, DE); Millie Irene Sterling (Wiesbaden, DE); Andrea Sanzone (Frankfurt, DE)

Assignee: THE BOEING COMPANY

G06N5/022

Abstract

Claims (59)

1 . A device comprising:

a memory; and

one or more processors coupled to the memory, the one or more processors configured to:

receive a data source that includes information associated with one or more airports;

determine, using a first machine-learning model, a particular classification of the data source and scheduling information associated with the data source;

allocate, using a second machine-learning model, particular information in the data source to a particular airport, the particular airport associated with a particular database, and the particular information scheduled to be descriptive of a feature of the particular airport;

select a third machine-learning model based on the particular classification determined by the first machine-learning model;

generate, using the third machine-learning model, an updated dataset based on the particular information, the updated dataset having a format similar to a current dataset in the particular database, the current dataset indicative of the feature of the particular airport;

update the current dataset in the particular database with the updated dataset based on the scheduling information in response to receipt of a user verification;

capture feedback data from the user verification, wherein the feedback data includes accuracy information indicating whether the updated dataset correctly identified changes to the feature of the particular airport;

generate training data by combining the feedback data with airport layout information extracted from airport layout images in the data source via natural language processing and image recognition of the airport layout images; and

retrain the selected third machine-learning model using the training data to improve classification accuracy for subsequent data sources of the particular classification related to the particular airport.

2 . The device of claim 1 , wherein the particular classification corresponds to a notice to airmen or an aeronautical information publication (AIP).

3 . The device of claim 1 , wherein the scheduling information indicates an effective date that the particular information will be descriptive of the feature of the particular airport.

4 . The device of claim 1 , wherein the particular database corresponds to an airport mapping database (AMDB).

5 . The device of claim 1 , wherein the one or more processors are configured to retrieve the current dataset from the particular database in response to allocating the particular information in the data source to the particular airport.

6 . The device of claim 1 , wherein the data source is received from an air navigation service provider (ANSP).

7 . The device of claim 1 , wherein, to allocate the particular information in the data source to the particular airport, the one or more processors are configured to:

convert one or more graphics associated with the particular information to text elements;

identify metadata in the text elements; and

allocate the particular information to the particular airport in response to a determination that the particular database includes equivalent metadata as the identified metadata.

8 . The device of claim 1 , wherein the one or more processors are configured to:

compare the updated dataset to the current dataset to identify a difference; and

generate an output indicating the difference, wherein the user verification is based on the output.

9 . The device of claim 1 , wherein the particular information corresponds to an aeronautical information publication (AIP) chart, AIP text, or both.

10 . The device of claim 1 , wherein the first machine-learning model is implemented using a first neural network, wherein the second machine-learning model is implemented using a second neural network, and wherein the third machine-learning model is implemented using a third neural network.

11 . A method comprising:

receiving, at one or more processors, a data source that includes information associated with one or more airports;

determining, using a first machine-learning model, a particular classification of the data source and scheduling information associated with the data source;

allocating, using a second machine-learning model, particular information in the data source to a particular airport, the particular airport associated with a particular database, and the particular information scheduled to be descriptive of a feature of the particular airport;

selecting a third machine-learning model based on the particular classification determined by the first machine-learning model;

generating, using the third machine-learning model, an updated dataset based on the particular information, the updated dataset having a format similar to a current dataset in the particular database, the current dataset indicative of the feature of the particular airport;

updating the current dataset in the particular database with the updated dataset based on the scheduling information in response to receiving a user verification;

capturing feedback data from the user verification, wherein the feedback data includes accuracy information indicating whether the updated dataset correctly identified changes to the feature of the particular airport;

generating training data by combining the feedback data with airport layout information extracted from airport layout images in the data source via natural language processing and image recognition of the airport layout images; and

retraining the selected third machine-learning model using the training data to improve classification accuracy for subsequent data sources of the particular classification related to the particular airport.

12 . The method of claim 11 , wherein the particular classification corresponds to a notice to airmen or an aeronautical information publication (AIP).

13 . The method of claim 11 , wherein the scheduling information indicates an effective date that the particular information will be descriptive of the feature of the particular airport.

14 . The method of claim 11 , wherein the particular database corresponds to an airport mapping database (AMDB).

15 . The method of claim 11 , further comprising retrieving the current dataset from the particular database in response to allocating the particular information in the data source to the particular airport.

16 . The method of claim 11 , wherein the data source is received from an air navigation service provider (ANSP).

17 . The method of claim 11 , wherein allocating the particular information in the data source to the particular airport comprises:

converting one or more graphics associated with the particular information to text elements;

identifying metadata in the text elements; and

allocating the particular information to the particular airport in response to a determination that the particular database includes equivalent metadata as the identified metadata.

18 . The method of claim 11 , further comprising:

comparing the updated dataset to the current dataset to identify a difference; and

generating an output indicating the difference, wherein the user verification is based on the output.

19 . The method of claim 11 , wherein the particular information corresponds to an aeronautical information publication (AIP) chart, AIP text, or both.

20 . A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to:

receive a data source that includes information associated with one or more airports;

determine, using a first machine-learning model, a particular classification of the data source and scheduling information associated with the data source;

allocate, using a second machine-learning model, particular information in the data source to a particular airport, the particular airport associated with a particular database, and the particular information scheduled to be descriptive of a feature of the particular airport;

select a third machine-learning model based on the particular classification determined by the first machine-learning model;

generate, using the third machine-learning model, an updated dataset based on the particular information, the updated dataset having a format similar to a current dataset in the particular database, the current dataset indicative of the feature of the particular airport;

update the current dataset in the particular database with the updated dataset based on the scheduling information in response to receipt of a user verification;

capture feedback data from the user verification, wherein the feedback data includes accuracy information indicating whether the updated dataset correctly identified changes to the feature of the particular airport;

generate training data by combining the feedback data with airport layout information extracted from airport layout images in the data source via natural language processing and image recognition of the airport layout images; and

retrain the selected third machine-learning model using the training data to improve classification accuracy for subsequent data sources of the particular classification related to the particular airport.

Assignments (1)

Continuity (1)

References Cited (11)