Workshop Fault Diagnosis Using Self-Learning Case Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current diagnostic systems for motor vehicle workshops face challenges in providing reliable and efficient troubleshooting due to the high variance in proposed diagnoses and test steps, especially with limited case data, leading to unreliable results and increased effort in maintaining knowledge bases across various vehicle brands and models.
Innovation Solution
A workshop diagnostic system that combines feedback data from a case database with author-created data, using an evaluation system to compare and match diagnostic data sets, generating modified records that improve troubleshooting accuracy by identifying typical faults and reducing the 'curse of dimensionality' through statistical analysis and self-learning capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If case-based databases are used for diagnostic support, then access to repair cases is improved, but the number of cases per vehicle is small leading to high variance in proposed causes
Solution Approach 1:
The patent combines case-based diagnostic data with author knowledge (predetermined diagnostic data sets created by experts) into a unified diagnostic system. This merging allows the system to leverage both the empirical repair cases and the structured expert knowledge, improving reliability when case numbers are low while maintaining access to actual repair experiences.
Solution Approach 2:
The system implements feedback mechanisms where diagnostic results from actual vehicle diagnoses are continuously fed back into the case database. This allows the system to learn from accumulated diagnostic experiences, improving the quality and reliability of proposed causes over time as more data becomes available for each vehicle type.
2Reliability
If author knowledge is used to build diagnostic systems, then reliability and variance of repair proposals is high, but time and cost for creating knowledge base increases
Solution Approach 1:
The patent stores predetermined diagnostic data sets created by authors in advance, organized by vehicle types and systems. This preliminary preparation of expert knowledge allows the system to provide reliable diagnostic suggestions immediately without requiring real-time expert intervention, reducing time loss while maintaining high reliability.
Solution Approach 2:
The system enables automatic diagnostic suggestions by storing and processing author knowledge in a structured format that can be queried and evaluated automatically. This self-service capability eliminates the need for continuous expert involvement in daily diagnostics, significantly reducing time and cost while maintaining reliable repair proposals.
3Adaptability or versatility
If manual authoring of diagnostic knowledge is performed, then expert experience is utilized, but the large number of brands and variants results in great effort
Solution Approach 1:
The patent segments diagnostic knowledge into vehicle-specific and system-specific data sets, allowing modular organization of expert knowledge. This segmentation enables independent maintenance and updates for different vehicle types and systems, reducing the overall complexity of maintaining knowledge bases across numerous brands and variants while maintaining comprehensive adaptability.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method for supporting a fault detecion on a technical object (2), in particular on a motor vehicle, having the following steps: receiving first diagnostic data from at least one object control device (4) and transmitting said data to a diagnostic server (12); generating at least one respective set of first diagnostic data from the transmitted first diagnostic data and assigning at least one first diagnostic result to the set of first diagnostic data; storing the first diagnostic data sets (181, 182) generated in the course of successive diagnoses in a first database (18); comparing the diagnostic data and/or the diagnostic result of each diagnostic data set of the first diagnostic data sets (181, 182) with the diagnostic data and/or the diagnostic result of specified second diagnostic data sets (201, 202) which are stored in a second database (20) and each of which contains a set of second diagnostic data and a second diagnostic result assigned to the set of second diagnostic data; assigning each first diagnostic data set (181, 182) to the specified second diagnostic data set (201, 202) with the greatest similarity to the respective first diagnostic data set (181, 182) according to the comparison; modifying each assigned second diagnostic data set (201, 202) using all of the first diagnostic data sets (181, 182) which are assigned to said second diagnostic data set (201, 202) in order to generate at least one third diagnostic data set (241, 242), said third diagnostic data set (241, 242) containing a respective set of third diagnostic data and at least one third diagnostic result assigned to the third diagnostic data set; and basing a support of the fault detection on the technical object (2) on at least one third diagnostic data set (241, 242).