Vehicle Sensor Cross-Validation for False Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor abnormality detection methods, such as those for lateral and vertical acceleration sensors, are prone to erroneous detection during steady circular turns or on sloped roads, leading to inaccurate sensor fault identification.
Innovation Solution
A sensor abnormality detection device that utilizes at least two sensors in a vehicle, comparing signal fluctuation ranges of these sensors based on vehicle speed thresholds to accurately determine sensor abnormalities, reducing erroneous detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor abnormality detection is performed by checking whether output value is outside normal range or by continuous monitoring, then detection capability is improved, but erroneous detection increases during steady circular turns or on sloped roads
Solution Approach 1:
The patent segments the detection process into multiple independent evaluation stages: initial abnormality detection based on output value range, continuous monitoring phase with time-based criteria, and cross-validation phase comparing multiple sensors. This segmentation allows the system to differentiate between temporary anomalies and genuine sensor failures, reducing false positives during steady-state operations like circular turns or sloped road travel.
Solution Approach 2:
The patent implements feedback mechanisms where detection results from one phase inform subsequent phases. The continuous monitoring phase provides feedback to confirm or refute initial abnormality detections, and cross-validation between multiple sensors provides additional feedback layers. This multi-stage feedback system ensures that erroneous detections during steady circular turns or sloped road travel are corrected before final abnormality determination.
2Productivity
If simple threshold-based detection is used, then detection speed is improved, but detection accuracy deteriorates in complex driving conditions
Solution Approach 1:
The patent applies preliminary threshold-based screening to quickly identify potential sensor abnormalities, then immediately follows with more sophisticated continuous monitoring and cross-validation procedures. This preliminary action approach maintains high detection speed for obvious failures while ensuring accuracy for borderline cases by preparing additional verification steps in advance.
Solution Approach 2:
The detection system dynamically adjusts its complexity based on operating conditions. During normal operation, simple threshold checks provide fast detection. When potential abnormalities are detected, the system dynamically transitions to more complex continuous monitoring and cross-validation modes, balancing detection speed and accuracy according to real-time needs.
Data Source
AI summary
A CAN input unit (traveling speed input unit) to which a traveling speed of a vehicle is to be input and sprung acceleration input units (a first sensor input unit and a second sensor input unit) to which signal fluctuation ranges (vibration levels) of sprung acceleration sensors are to be input are included. An ECU includes a first determination unit which determines whether a traveling speed provided by the CAN input unit is equal to or more than a first threshold value, and a signal fluctuation range of one of a sprung acceleration sensor value of the sprung acceleration input unit serving as the first sensor input unit and a sprung acceleration sensor value of the sprung acceleration input unit serving as the second sensor input unit is smaller than a first fluctuation range and smaller than a signal fluctuation range of another of the first and second sensor input values.