Michael Vössing

Background

Estimating the health of vehicle components helps to make the automotive experience safer and more sustainable. Since modern vehicles come with an enormous amount of operating data, data-based approaches such as machine learning (ML) are suitable estimation techniques. While research provided a path to deployed ML-based health estimation in e.g. manufacturing plants, ML-based health estimation in the automotive industry is scarce due to a significantly more challenging environment. Hence, there remains a large potential for ML applications.

Research Goal

In this thesis, we want to study how diagnostic data and feedback from maintenance workshops can be used to estimate the health of vehicle chassis components. This way we make vehicles more reliable and revolutionize the maintenance process. The complexity of estimation is challenging due to the large variety of car lines and the continuous changing data base (e.g., more mileage). That's why it is desired to continuously learn and analyze the risk and uncertainty of predictions. A real world diagnostic data set is provided for this research.

Working on this thesis you will:

- apply your theoretical knowledge to a practical use case in the context of predictive maintenance.

- use ML algorithms to estimate the health of vehicle components by anomaly detection.

- further develop and evaluate existing estimation methods for health state estimation and failure forecasting.

- become an expert in applied data analytics and ML.

- work directly with domain experts to improve the performance of the system.

We look forward to receiving your application because you...

- have solid programming skills in Python 

- have experience with state-of-the-art ML  and data analytic methods---or are willing to learn quickly

- team orientated and highly motivated

- want to gain first-hand experience solving real-world problems

Details

- Start: As soon as possible 

- Duration: 6 Month 

- Salary: Yes

- Location: CARIAD, Remote or nearby Weissach