Aircraft Anomaly Detection Using Big Data Clustering and Statistical Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computer analytic systems for detecting aircraft anomalies are less accurate due to their focus on smaller subsets of flight data, rather than the massive set of data generated by aircraft, leading to inefficiencies in analyzing big data and identifying issues effectively.
Innovation Solution
A method using a big data analytic computing device that clusters aircraft flight data into groups, determines the distance between these groups and baseline data, and executes statistical model analysis to detect anomalies, incorporating modules for Euclidean, dynamic time warping, and correlation-based distance calculations, along with regression, Markovian, and compression models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional data processing applications are used to analyze aircraft flight data, then the system complexity is low, but the analysis accuracy and anomaly detection effectiveness deteriorate due to inadequate handling of big data
Solution Approach 1:
The patent segments the massive flight data into smaller directed subsets that can be processed by traditional applications, while using a hybrid approach that combines big data analytics for overall patterns with traditional methods for detailed analysis, thus improving detection accuracy without requiring complete system replacement
Solution Approach 2:
The patent introduces an intermediary layer that bridges big data analytics and traditional data processing applications. This intermediary processes and prepares data subsets, enabling traditional applications to work effectively with big data without requiring full system complexity, thus resolving the contradiction between accuracy improvement and complexity increase
2Device complexity
If smaller directed subsets of flight data are analyzed, then the data processing complexity is reduced, but the anomaly identification accuracy deteriorates
Solution Approach 1:
The patent divides the massive flight data into smaller directed subsets that are manageable by traditional applications, allowing complex data to be processed in smaller chunks without losing the ability to identify anomalies through strategic selection and analysis of these subsets
Solution Approach 2:
The patent analyzes specific subsets of data that are most relevant to anomaly detection rather than processing the entire dataset, achieving effective anomaly identification with reduced computational complexity by focusing on critical portions of the data
Data Source
AI summary
Methods, devices, and non-transitory computer readable media that detect an anomaly in an aircraft include obtaining aircraft flight data from multiple aircraft sensor devices. The obtained aircraft flight data is clustered into two or more data groups. A distance between the clustered aircraft flight data in at least one of the two or more data groups associated with a part of the aircraft and stored baseline flight data for the part of the aircraft is determined. A statistical model analysis is executed on the determined distance to detect any anomaly with the part of the aircraft.