In the automotive industry, more and more assistance systems are being introduced to support the driver, e.g. parking assistance or automatic parking steering systems. These systems exploit data from safety-relevant sensors. The goal of this thesis was to find new methods to detect faulty sensors at an early stage in production.
Techniques from automated machine learning were used in this thesis to derive classification models for sensors. These models not only allow to identify defective sensors, but also help to identify the type of error. In the thesis, the models were applied to real-world data. The results show the potential to improve production processes as well as increase customer satisfaction by reducing the number of customer returns. |