UniBwM » EIT3_2 » Chair of Signal Processing » Home » News » 2014 and older » Information Engineering at the automotive "Human-Machine-Interface" – PhD Graduation awarded by “summa cum laude”

Information Engineering at the automotive "Human-Machine-Interface" – PhD Graduation awarded by “summa cum laude”

Information Engineering at the automotive "Human-Machine-Interface" – PhD Graduation awarded by “summa cum laude”


March, 8th, 2010:   Presenting a PhD thesis entitled “Adaptive Information Flow Control – Recognition and Prediction of Factors contributing to Driver’s Stress” at the Institute of Communication Engineering (Chair 'Information Engineering'), Dipl.-Inform. Michael Dambier (external PhD student) received his PhD degree, awarded 'summa cum laude' by the the faculty board of Prof. Hillermeier (chairman), Prof. Bauch, Prof. Wolf, and Prof. Appel. The PhD project was conducted in collaboration with the Corporate Sector Research and Advance Engineering at the Robert Bosch GmbH Stuttgart. After the examination procedure, Michael Dambier received the traditional 'Doktorhut' designed in a very creative way by his colleagues from the institute and the company.

The Focus of the Project: Driver's Stress and Strain

The work of Michael Dambier addresses the factors contributing to driver’s stress and strain. First, stress factors are identified by a comprehensive literature review and structured into the topics 'road type', 'road characteristic', 'vehicle environment', 'weather and resulting road condition', 'driving maneuvers', and 'secondary tasks'. On this base, the recognition and prediction of driving maneuvers and the estimation of the traffic density in the vicinity of a vehicle (as a possible application of driving maneuver recognition) are designed. Maneuver specific features for classification are derived on the basis of analytical decomposition of each driving maneuver as well as on an extensive maneuver related statistical data analysis.
Using these features, an approach combining the methods 'decision tree' and 'Random Forest' is realized for the recognition of the single driving maneuvers. The final classification also considers the average maneuver durations and maneuver transition probabilities. Additionally, a maneuver prediction algorithm is shown using probabilities of maneuver sequences of different length. Based on the assumption that traffic density has a strong impact on the set of executed driving maneuvers, a traffic density estimation can be performed using driving maneuver recognition. Two comprehensive real drive experiments were conducted on public roads to obtain the data necessary to conduct this PhD project.

What is the final contribution of the project?

Besides a comprehensive structuring of driver’s stress factors which allows more systematic concepts in future work, a driving maneuver classification system is presented being capable for in-vehicle online application and showing excellent classification results compared to reported results in literature. Furthermore, a novel concept for traffic density estimation in the vicinity of a vehicle on the basis of the developed driving maneuver recognition system is now available for future efforts to provide another assistance tool to the driver. This yields a valuable contribution to the design of human-machine-interfaces in automotive applications.