Exploring Lane Change Style Recognition through Analysis of Latent Driving Behavior

In the context of high-speed mixed traffic, a challenge arises in the recognition accuracy of lane change driving styles due to the oversight of latent driving behavior information by existing models. To address this issue, we propose a lane change style recognition model based on Long Short-Term Memory (LSTM) networks. Initially, by emulating the attention mechanism of human drivers towards surrounding vehicles, we analyze driving behaviors. Subsequently, the model is trained using the vehicle's motion state data and interaction features with surrounding vehicles as inputs, categorizing lane change styles into conservative, normal, and aggressive. Finally, the effectiveness of the model is validated using the NGSIM dataset. The results demonstrate that the improved lane change style recognition model achieves an accuracy of 94.17%, confirming its high precision in identifying vehicle lane change styles.