Abstract: In order to prevent traffic accidents caused by driver distraction, a method was developed for distraction detection by using the radial basis function neural network model. The driving behavior during the lane change process was analyzed through driving simulation experiments. On the basis of the data from the experiments, effects of driver's three states, normal driving, hand-held answering phone and hand-free answering phone on the lane changing behavior were studied. An RBF neural network model was built based on orthogonal least square(OLS)method to monitor whether the driver is driving distracted. The results show that driving distraction during the lane change process has a significant influence on 6 driving performance parameters such as vehicle's longitudinal velocity, lateral velocity, lateral acceleration, steering angle, steering wheel velocity and opening degree of accelerator, and the average detection accuracy of the model reaches 88.7%, which could accurately identify the distracted state of the driver and provide theoretical support for the prevention of distraction accidents.

Key words: driving safety, distraction detection, radial basis function(RBF), lane change, driving simulation