Artificial Intelligence in Technical Applications

Artificial intelligence (AI) is used in all areas of the institute.
In addition to intelligent test planning using machine learning, the Stuttgart Reliability Research Center is also investigating how the remaining service life of components can be calculated based on existing data using AI methods.
To avoid unexpected failures of electronic systems, error precursors in data sets should be detected, thereby enabling predictive maintenance. AI algorithms for anomaly detection are used here to identify error precursors as deviations from normal operating conditions without requiring a specific definition of the data points being sought. In addition, the detectability of failures in electronic components in a circuit is evaluated using classification algorithms. In order to make efficient decisions in the development of Prognostics and Health Management (PHM) algorithms, hybrid anomaly detection algorithms compare different types of expert knowledge in terms of their usefulness for error detection. In addition, the generation of synthetic error data using generative AI algorithms and error simulations of electronic circuits is being investigated in order to compensate for missing failure data from new or highly reliable systems.
AI is also used in the field of autonomous driving to monitor the structural integrity of vehicles during collisions. AI is used to analyze the level of damage and derive recommendations for operational action. AI is also used in the context of autonomous train driving. The aim is to support and relieve the train driver and optimize their driving behavior in terms of safety, wear and tear, energy efficiency, and comfort.
In the research area Drive Technology, AI is used to detect gear damage at a very early stage during operation. Sparse autoencoders and other machine learning methods are used. The implementation of adaptive operating strategies to increase the service life of gearboxes through intelligent torque control is also being investigated.
In the field of Sealing Technology, AI-based image analysis is primarily used to detect and classify different types of damage.