Sensor Fault Detection via Online Correlation Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and diagnosing sensor faults in autonomous systems, such as robots and UAVs, are inefficient in quickly identifying faults like 'stuck' and 'drift' conditions, which can lead to mission failures due to their reliance on mathematical models or data-driven approaches that require extensive data processing and lack accurate fault isolation.
Innovation Solution
A method that uses a structural model to represent sensor dependencies on hardware components, employing online correlation tracking and Pearson Correlation Coefficient calculations to identify suspicious patterns and isolate faulty sensors, allowing for real-time fault detection and diagnosis without requiring extensive labeled data or mathematical representations of component behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If mathematical models or data-driven approaches are used for fault detection, then fault detection capability is provided, but false alarm rates increase and detection speed decreases
Solution Approach 1:
The system performs preliminary actions by continuously tracking correlations between sensors and hardware components online, maintaining up-to-date correlation information before faults occur. This allows the system to quickly identify suspicious patterns when faults happen, reducing detection time while maintaining accuracy through pre-established correlation knowledge.
Solution Approach 2:
The patent segments the fault detection process into distinct stages: correlation tracking between sensors and hardware components, suspicious pattern identification, and fault isolation. This segmentation allows each stage to be optimized independently, improving overall detection speed and reducing false alarms by focusing computational resources on specific tasks.
2Reliability
If mathematical models or data-driven approaches are used for fault detection, then fault detection capability is provided, but false alarm rates increase
Solution Approach 1:
The patent introduces correlation coefficients as an intermediary measure between sensor readings and hardware component states. Instead of directly applying complex mathematical models to raw sensor data, the system uses correlation coefficients to mediate the relationship, providing a more robust indicator that reduces false alarms while maintaining detection accuracy.
Solution Approach 2:
The system implements feedback by continuously monitoring correlation patterns and using this information to adjust fault detection decisions. The online tracking of correlations provides ongoing feedback about system health, allowing the system to distinguish between normal variations and actual faults, thereby reducing false alarm rates.
3Measurement precision
If extensive data processing is performed for fault detection, then detection capability is improved, but system complexity and computational requirements increase
Solution Approach 1:
The patent extracts only the essential correlation information between sensors and hardware components, rather than processing all available sensor data. By taking out and focusing on the most relevant correlation metrics, the system achieves accurate fault detection with reduced computational complexity and simpler system requirements.
4Measurement precision
If correlation tracking is performed online between sensors and hardware components, then fault isolation accuracy is improved, but computational load increases
Solution Approach 1:
The system performs partial correlation tracking by focusing computational resources on tracking correlations for sensors and components that are most relevant to current operation. Rather than continuously computing all possible correlations, the system applies correlation tracking selectively, reducing computational energy consumption while maintaining fault isolation accuracy for critical components.
Data Source
AI summary
A method for detecting and diagnosing sensor faults in an autonomous system that includes sensors and hardware components, according to which sensors are related to hardware components and correlations between data readings are recognized online and correlation between sensors is determined. Predefined suspicious patterns are identified by online and continuously tracking the data readings from each sensor and detecting correlation breaks over time. The readings from sensors that match at least one of the patterns are marked as uncertain. For each online reading of the sensors, whenever sensors that used to be correlated show a different behavior, reporting that the reading indicates a fault. Upon identifying fault detection, diagnosing which of the internal components or sensors caused the fault, based on a function that returns the state of the sensor which is associated with the fault detection.