Focus:

  • Machine Learning / Stochastic Processes / Nonparametric Bayesian
  • Vehicle state prediction for longitudinal and lateral dynamics
  • Prediction of driver depended energy demand
  • Prediction of the driving route-trajectory

Summary:

Decisive for the vehicle state prediction as well as advanced driver assistance systems (ADAS) in the automotive industry is the modelling of the overall system "driver - vehicle - driving environment". The figure demonstrates the three different models between which diverging interactions and influencing variables exist.

The dynamic vehicle condition of a vehicle is influenced by driver behavior and external conditions. In this system, the driver is responsible for vehicle driving, depending on environmental influences, EGO vehicle data and psychological needs.

In order to influence the vehicle condition predictively and efficiently, the research project is concerned with an online calculation of the driving behavior with regard to longitudinal and lateral dynamics, as well as the speed behavior on the basis of

  • Predictive environmental data (e.g. map, sensor and weather data, front vehicle information, online information, Car2X Communication, etc.),
  • Ego vehicle data (e.g. load, engine, vehicle resistance, etc.)
  • and different driving behavior and driver types

Algorithms and functions are developed for this purpose, which are then verified and evaluated on the basis of test drives and studies.

The detailed prediction of the driving behavior can be used to optimize the vehicle condition, for example the operating strategy, which, among other things, leads to economical and efficient operation. In addition, assistance functions receive precise information to support the driver.
 

Contact

Head of the Institute for Innovative Mobility (IIMo) Research Professorship E-Mobility and Learning Systems
Prof. Dr.-Ing. Christian Endisch
Phone: +49 841 9348-5171
Room: A116
E-Mail: