Sürücü davranış sinyalleri ile sürücü statüsü tanıma

Driving behavior signals differ in how and under which conditions the driver use vehicle control units, such as pedals, driving wheel, etc. In this study we investigate how the driving behavior signals differ among drivers and among different driving tasks. Statistically significant clues of these investigations are used to define driver and driving status models. Experimental results over the UYANIK database are presented. Driver identification over 23 drives achieves 57.39% identification rate with the fusion of gas and brake pedal pressure classifiers. Driver identification system with reduced number of drivers suits better to real-life scenario. 85.21% of identification rate is achieved among 3 drivers. Driver status identification over 10 drivers with task and no-task classes yields a promising 79.13% identification rate.Driving behavior is strongly related to past movements of drivers. In this thesis we aim to predict driving behavior for warning drivers about future incidents and decreasing car accidents caused by human factors. The proposed method is concerned with past samples of behavior signals and we use Hidden Markov Models to model driving behavior. Earlier findings have shown us that we can predict driving behavior with encouraging results in both driver dependent and independent experiments. The experimental results also show that distractive conditions have a certain effect on driving behavior as the prediction errors are significantly increasing in these conditions. Road conditions are also influential on driving.