Systems and methods for aircraft management
AssigneeAermetric Technology Group, Inc.
InventorsSean BUSHKOV, JR., Derek WATERMAN
Life of this patent
Estimating the term…
Abstract
In an aspect, a method for monitoring recurrent aircraft faults is disclosed. The method comprises (a) retrieving a plurality of data entries from an aircraft. The plurality of data entries may comprise semantic information about a plurality of recurrent aircraft faults. The method also comprises (b) using a trained natural language processing algorithm, establishing a plurality of pairwise relations of the plurality of data entries, a pairwise relation corresponding to a recurrent fault of the plurality of recurrent aircraft faults. The method also comprises (c) generating one or more alerts from the plurality of pairwise relations. Finally, the method comprises (d) presenting the one or more alerts in a user interface.
Claims20
A method for monitoring recurrent aircraft faults, comprising:
- (a) retrieving a plurality of data entries from an aircraft, the plurality of data entries comprising semantic information about a plurality of recurrent aircraft faults;
- (b) using a trained machine learning algorithm, establishing a plurality of pairwise relations of the plurality of data entries, a pairwise relation corresponding to a recurrent fault of the plurality of recurrent aircraft faults;
- (c) generating one or more alerts from the plurality of pairwise relations;
- and
- (d) presenting the one or more alerts in a user interface.
A computer-implemented method of training a machine learning model for detecting recurrent aircraft faults, comprising:
- (a) receiving data from a plurality of sources;
- (b) applying one or more pre-processing operations to the data, wherein the one or more pre-processing operations comprises one or more of:
- combining the data, compressing the data, parsing the data, or standardizing the data to create pre-processed data;
- (c) creating a training set comprising a plurality of pairwise relations associated with the pre-processed data;
- and
- (d) training the machine learning model to link a pair of the plurality of pairwise relations.
A system for monitoring recurrent aircraft faults, comprising:
- one or more computer processors individually or collectively programmed to:
- (a) retrieve a plurality of data entries from an aircraft, the plurality of data entries comprising semantic information about a plurality of recurrent aircraft faults;
- (b) use a trained machine learning algorithm, establishing a plurality of pairwise relations of the plurality of data entries, a pairwise relation corresponding to a recurrent fault of the plurality of recurrent aircraft faults;
- (c) generate one or more alerts from the plurality of pairwise relations;
- and
- (d) present the one or more alerts in a user interface.
A non-transitory computer-readable medium comprising machine-executable code that, upon execution by one or more computer processors, implements a method for monitoring recurring aircraft faults, the method comprising:
- (a) retrieving a plurality of data entries from an aircraft, the plurality of data entries comprising semantic information about a plurality of recurrent aircraft faults;
- (b) using a trained machine learning algorithm, establishing a plurality of pairwise relations of the plurality of data entries, a pairwise relation corresponding to a recurrent fault of the plurality of recurrent aircraft faults;
- (c) generating one or more alerts from the plurality of pairwise relations;
- and
- (d) presenting the one or more alerts in a user interface.
Dependencies are inferred from the claim text.