Study on Individual Driving Style Based on Driving Behavior Characteristics

摘要

61 experienced drivers were recruited to implement driving simulator tests. Participants were asked to drive along a 21 kilometer experiment track, during which the vehicle driving state was recorded. Small traffic volumes were designed in the simulator environment to achieve a vehicle's free flow speed. The frequency of sudden braking and acceleration per kilometer was extracted as the feature parameter describing the individual's driving style. Then, all drivers were clustered into five clusters with the aid of K-means, and the five clusters represent if the driver has a normal, aggressive, very aggressive, cautious, and very cautious driving style. It showed that the aggressive driver had a large frequency of sudden braking and acceleration, while the cautious driver's had a small frequency of these behaviors. However, among all drivers, aggressive drivers did not have the maximum average driving speed along the experiment, but rather the normal and cautious driver had the large average driving speed.