Redundant Vehicle Subsystem Monitoring for Anomaly Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Detecting anomalies and isolating faults in complex systems, such as aircraft, is challenging due to limited sensor data, low sample data, and the difficulty in characterizing nominal behavior under varying operating conditions.
Innovation Solution
The method involves identifying redundant sub-systems, forming combinations of these sub-systems, selecting the combination with the least variability, computing nominal values for parameters, and detecting anomalies based on these nominal values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional anomaly detection methods are used, then detection capability is limited, but computational complexity and data requirements increase
Solution Approach 1:
The system segments the complex system into multiple redundant subsystems, each monitored independently. By dividing the monitoring task across subsystems and using ensemble methods, the computational burden is distributed while maintaining detection accuracy. The segmentation of feature extraction into multiple independent feature sets further reduces the complexity of analyzing each individual feature.
Solution Approach 2:
The system merges multiple feature sets from different subsystems to form a comprehensive anomaly detection model. By combining features from redundant subsystems and using ensemble learning methods, the system achieves higher detection accuracy while the merging process itself helps filter out noise and reduce the impact of individual feature variations.
2Loss of information
If comprehensive sensor data collection is implemented, then measurement coverage improves, but data processing complexity increases
Solution Approach 1:
The system extracts only the most relevant features from comprehensive sensor data using feature selection techniques. By taking out and focusing on critical features rather than processing all available sensor data, the system maintains complete information coverage while significantly reducing processing complexity. The feature extraction process identifies and isolates key indicators of system health.
Solution Approach 2:
The system performs preliminary feature extraction and selection before main anomaly detection processing. By pre-processing the sensor data to identify and extract relevant features in advance, the system reduces the dimensionality of data that needs to be processed during runtime, thereby lowering computational complexity while preserving all necessary information for accurate detection.
3Measurement precision
If nominal behavior characterization is performed under varied operating conditions, then detection accuracy improves, but computational requirements increase
Solution Approach 1:
The system performs partial characterization of nominal behavior by focusing on the most critical operating conditions and features rather than attempting to model all possible variations. This partial action approach achieves sufficient detection accuracy for safety-critical applications while consuming less computational energy than comprehensive modeling would require.
4Measurement precision
If redundant subsystems are monitored individually, then fault isolation precision improves, but processing time increases
Solution Approach 1:
The system replaces traditional sequential mechanical-style inspection of each subsystem with parallel computational analysis using ensemble methods. Multiple subsystems are analyzed simultaneously through parallel processing and ensemble algorithms, maintaining precise fault isolation capability while dramatically reducing processing time compared to sequential individual monitoring.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Techniques for anomaly detection are disclosed. These techniques include identifying a plurality of three or more redundant sub-systems in a system. The techniques further include forming a plurality of combinations of the redundant sub-systems, each combination relating to a subset of the plurality of redundant sub-systems. The techniques further include identifying a first combination with the least variability among sub-systems, from among the plurality of combinations of redundant sub-systems. The techniques further include computing one or more nominal values for one or more parameters of a first sub-system, of the plurality of sub-systems, using the first combination, and detecting an anomaly in the first sub-system based on the computed one or more nominal values.