Abstract: In view of the correlation between the traffic density and the driver’s pedaling behavior, this paper proposes a multi-level driving style identification method that considers the influence of traffic density based on the coupling relationship between traffic density and driving style identification. Firstly, based on the simulation driving experiment platform, the data of pedal signals and speed signals of drivers under different traffic densities are collected and the characteristic parameters of different driving styles are extracted. Then, principal component analysis is used to obtain the comprehensive characteristic parameters of driving styles under the influence of different traffic densities, and the hybrid algorithm of subtractive clustering and K-means clustering is applied to classify the driving styles and modify the characteristic parameters of different driving styles on this basis. Finally, random forest algorithm is used to identify and verify the driving styles. The results show that under the influence of different traffic densities, the method proposed in this paper has high accuracy for driving styles identification, which lays a foundation for further optimization of energy management strategy for hybrid electric vehicles.

Key words: driving style, traffic density, coupling, modification of characteristic parameters, multi-level identification method