Driving Pattern Recognition Algorithm Using Fast Fourier Transform

Recognition of driving patterns is widely used in online applications such as advanced driver assistance systems and energy management of electric vehicles or fleet management, vehicle security, insurance policies, vehicle to grid to detect driving conditions and driver behavior. This article presents a driving pattern recognition method using Fast Fourier Transform that can be used offline or in online applications that allow a resolution of around 1 km for the classification of energy consumption. Within the scope of work, driving pattern categories are defined considering the road types in Turkey and the driving events that are characterized by the driving style of the users such as acceleration, braking, stopping or going at a steady speed by examining the distinctive features of the speed data in time and Fourier domain. Then, the performance of the pattern recognition algorithm is evaluated on a control group showing different driving styles.