Prediction of Erroneous Car Driver Behavior
Frédéric Vanderhaegen, Philippe Polet · 15th World Congress on Intelligent Transport Systems and ITS America's 2008 Annual MeetingITS AmericaERTICOITS JapanTransCore · 2008
This paper focuses on the use of the Benefit-Cost-Deficit (BCD) model as a framework for input data vectors of redundant prediction tools. A case-based reasoning system and a neural network system are developed and applied to predict particular intentional human errors: traffic violations called barrier removals made the car drivers. Two groups of input data are compared: a group containing the data from the car and from the human actions and a group containing the translation of these data in terms of BCD parameters. Results obtained by the redundant prediction tools using both groups of input data are discussed.