Seminar Friday November 4
Probabilistic classifiers for turn and lane change detection
Joyce MacDuff
For the future of automated driving, it will be desirable that vehicles will be able to react to their surroundings. These
vehicles will therefore have to be able to automatically detect traffic events. In order to develop such algorithms, large amounts
of data will have to be dissected into different traffic events. TNO have developed algorithms for this purpose, regarding turn and
lane change detection using only basic in car sensors. This thesis is about the optimization of these algorithms. The uncertainty
regarding these detectors is expressed using probabilistic classifiers.
In the presentation the motivation for such detectors, the data set that is used, the methods that are used to optimize these
algorithms and the methods used to express the uncertainty will be discussed.