Anomality Candidate Analysis for Vehicle Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing failure diagnosis apparatuses for vehicles struggle to early detect unexpected anomalies, relying solely on information acquisition which is insufficient for prompt confirmation.
Innovation Solution
An anomality candidate information analysis apparatus that stores and analyzes monitoring object, environment, and operator information to determine the dependence degree of anomality candidate information, enabling early detection and cause specification of unexpected anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If only failure information is acquired when a failure is detected, then the system maintains simple information collection, but it cannot early detect unexpected anomalies
Solution Approach 1:
The system collects monitoring object information, environment information, and operator information in advance before failures occur. This preliminary data collection enables the system to analyze dependencies and detect unexpected anomalies early, rather than waiting for failures to manifest. The storage unit saves these diverse information types associated with each other for future analysis.
Solution Approach 2:
The information collection system is divided into three distinct segments: monitoring object information (vehicle state), environment information (surrounding conditions), and operator information (driver behavior). This segmentation allows the system to analyze each information type's contribution to anomalies independently while understanding their combined effects through dependency analysis.
2Loss of time
If comprehensive information is collected and analyzed, then early anomaly detection is enabled, but the analysis complexity increases
Solution Approach 1:
The analysis unit transforms multiple information parameters (monitoring object, environment, operator) into a unified dependency degree metric. By changing the parameter representation from raw diverse data to a standardized dependency measure, the system simplifies the analysis process while maintaining comprehensive anomaly detection capabilities across different information types.
Solution Approach 2:
The dependency degree analysis acts as an intermediary mechanism that bridges the gap between diverse information sources and anomaly detection. Instead of directly analyzing complex multi-source data, the system uses dependency degree as an intermediate metric to quantify relationships between information types and anomaly occurrences, simplifying the overall analysis process.
3Measurement precision
If dependency analysis is performed on multiple information types, then cause specification accuracy improves, but computational requirements increase
Solution Approach 1:
The analysis unit extracts only the essential dependency relationships between information types and anomaly occurrences, rather than processing all possible combinations of data. By taking out and focusing on the most relevant dependencies, the system achieves high cause specification accuracy while minimizing unnecessary computational energy consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In order to provide an anomality candidate information analysis apparatus capable of early confirming that an unexpected anomality occurs in a monitoring object such as a vehicle, there is provided an anomality candidate information analysis apparatus for analyzing anomality candidate information of a vehicle that is a monitoring object. The anomality candidate information analysis apparatus includes a storage unit (30) and an analysis unit (40). The storage unit (30) stores monitoring object information regarding the vehicle, environment information regarding an environment around the vehicle, operator information regarding an operator of the vehicle, and anomality candidate information detected in the vehicle in association with each other. The analysis unit (40) extracts the anomality candidate information associated with the monitoring object information, the environment information, and the operator information, and analyzes a dependence degree of the anomality candidate information to the environment information and the operator information.