Driver identification using variance of the acceleration data

This paper presents a novel methodology for driver identification using hidden characteristics in the variance of acceleration data of the target vehicles. The proposed method is based on the use of raw acceleration data of moving vehicles collected from mobile devices such as smartphone which provides an easy access solution in comparison to existing approaches based on bio-sensors, cameras and steering wheel movements. Results from the analysis showed that the proposed driver identification method can tell the relationship of the acceleration data of each driver and lead to guidelines to distinguish the driving habits of each motorist.