NorCom Information Technology GmbH & Co. KGaA has successfully completed an AI project for prediction at a German automobile manufacturer that has been running since the beginning of the year. NorCom supported the customer in making predictions for error codes detected during driving based on an AI model. The project is thematically assigned to the larger field of root cause analysis. In a planned follow-up project, the knowledge gained will serve as a basis for the development of several AI apps on various aspects of root cause analysis.
Gradual implementation of the Root Cause Analysis
Root Cause Analysis determines how, why, and when a problematic event occurs. The aim is to be able to systematically avoid a problem that occurs in the future by analyzing the causes, instead of solving it afterwards. The results of the root cause analysis also provide important information to determine the underlying causes of errors and to be able to rectify them in a cost-efficient and customer-friendly manner.
As part of series support, error codes that occur in all vehicles sold are monitored and evaluated, provided the owners have given their consent. NorCom supports this error code analysis with the prediction. The prediction is based on an AI training model and examines relationships between error codes and internal and external factors that occur in parallel.
Load collectives, endurance run data and error environment data flow into the training data, which are calculated directly at the time of driving, as well as the error codes that have accumulated. The error codes can be related to external circumstances such as driving frequency, location, weather conditions or driving speed. But also instruments and sensors inside the car, as well as their interactions, are included in the evaluation.
"The customer's task was predestined for the sophisticated data analysis that we can carry out with DaSense," comments Dr. Tobias Abthoff the project. “Our solution processes huge amounts of raw data that, due to their heterogeneity and scope, cannot be viewed and interpreted by humans. The evaluation by artificial intelligence and the creation of correlations gives the data new meaning. Connections that are not obvious are essential for the further optimization of vehicles.”
Outlook
In a planned follow-up project, the knowledge gained from the prediction should serve as a basis for the development of further AI apps for root cause analysis, which can then be used productively by the customer.
Gradual implementation of the Root Cause Analysis
Root Cause Analysis determines how, why, and when a problematic event occurs. The aim is to be able to systematically avoid a problem that occurs in the future by analyzing the causes, instead of solving it afterwards. The results of the root cause analysis also provide important information to determine the underlying causes of errors and to be able to rectify them in a cost-efficient and customer-friendly manner.
As part of series support, error codes that occur in all vehicles sold are monitored and evaluated, provided the owners have given their consent. NorCom supports this error code analysis with the prediction. The prediction is based on an AI training model and examines relationships between error codes and internal and external factors that occur in parallel.
Load collectives, endurance run data and error environment data flow into the training data, which are calculated directly at the time of driving, as well as the error codes that have accumulated. The error codes can be related to external circumstances such as driving frequency, location, weather conditions or driving speed. But also instruments and sensors inside the car, as well as their interactions, are included in the evaluation.
"The customer's task was predestined for the sophisticated data analysis that we can carry out with DaSense," comments Dr. Tobias Abthoff the project. “Our solution processes huge amounts of raw data that, due to their heterogeneity and scope, cannot be viewed and interpreted by humans. The evaluation by artificial intelligence and the creation of correlations gives the data new meaning. Connections that are not obvious are essential for the further optimization of vehicles.”
Outlook
In a planned follow-up project, the knowledge gained from the prediction should serve as a basis for the development of further AI apps for root cause analysis, which can then be used productively by the customer.
