Intelligent Vehicle Fault Diagnosis Using De-Noised Sensor Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fault diagnosis methods for intelligent vehicles are inadequate in accurately diagnosing faults in complex systems and positioning faults in subsystems using sensor data, as they rely on traditional industrial methods that are not systematically effective.
Innovation Solution
A method involving establishing a system model for intelligent vehicles, acquiring and de-noising system operation data, performing feature extraction and screening for fatal sensor faults, and comparing system state data with thresholds to determine actuator faults, utilizing discrete wavelet transform and sliding window methods for accurate fault detection and positioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional fault diagnosis methods (model-based, signal-based, data-driven) are used for intelligent vehicles, then the diagnostic process can be performed, but the accuracy of fault detection and positioning in complex systems is insufficient
Solution Approach 1:
The patent segments the fault diagnosis process into distinct modules: data acquisition module, data cleaning module, feature extraction module, fault judgment module, and fault positioning module. Each module handles specific aspects of the diagnosis process, improving overall accuracy through specialized processing at each stage.
Solution Approach 2:
The patent introduces discrete wavelet transform as an intermediary technique between raw sensor data and fault detection. The wavelet transform decomposes complex sensor signals into manageable frequency components, enabling more accurate feature extraction and fault identification in the intermediate processing stage.
2Measurement precision
If sensor data is used directly for fault positioning, then the process is simple, but the accuracy of fault positioning in subsystems is insufficient due to noise and complex system interactions
Solution Approach 1:
The patent performs preliminary data cleaning and feature extraction before fault positioning. Sensor data undergoes cleaning to remove noise and outliers, followed by feature extraction to identify relevant characteristics. This preliminary processing prepares the data for accurate fault positioning while managing complexity through structured preprocessing steps.
Solution Approach 2:
The patent replaces direct mechanical analysis of sensor data with signal processing techniques, specifically discrete wavelet transform. This substitution transforms the complex mechanical system analysis into a mathematical transformation process, enabling more accurate fault positioning through frequency-domain analysis rather than direct time-domain inspection.
3Measurement precision
If comprehensive system operation data is collected and processed, then fault detection accuracy improves, but the processing time and computational resources increase
Solution Approach 1:
The patent extracts only the most relevant features from comprehensive sensor data through the feature extraction module. Instead of processing all raw data, the system identifies and extracts key characteristics using discrete wavelet transform, reducing the data volume that requires further processing while maintaining diagnostic accuracy.
Solution Approach 2:
The patent applies partial processing to the sensor data by focusing on specific frequency bands and features that are most indicative of faults. The discrete wavelet transform decomposes data into different resolution levels, and the system processes only the relevant levels and features, avoiding unnecessary computation on less important data while maintaining diagnostic accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables effective detection and positioning of fatal sensor faults and state abnormalities in intelligent vehicles, improving the accuracy and efficiency of fault diagnosis by using a model-based system with de-noised and reconstructed data.
Implementation Method 1
The feature extraction is performed for the system operation data of the intelligent vehicle in the normal running state at different scales using discrete wavelet transform (DWT)
Implementation Method 2
a sliding window method is adopted for the DWT
Data Source
AI summary
A model of a system of an intelligent vehicle is trained and optimized using system operation data of the intelligent vehicle in a normal running state. The system operation data of the intelligent vehicle in a running state is collected in real time. Sensor data of the system operation data is de-noised, and feature extraction and screening are performed for a fatal sensor fault to reconstruct the system operation data. The reconstructed system operation data is inputted into the trained model to output system state data of the intelligent vehicle in the running state. The system state data is compared with a set threshold. If the system state data exceeds the set threshold, an actuator corresponding to the system state data is determined to have a fault. In addition, a system for a fault diagnosis of the intelligent vehicle is further provided.