Research on Abnormal Driving Behavior Identification Methods for Intelligent Network Connection

In view of the frequent traffic accidents in daily life, a method for identifying abnormal driving behavior for intelligent network connection was studied. Through the organic combination of the self-attention mechanism and the deep residual network model, the model parameters are optimized and trained to realize the identification of abnormal driving behaviors. Through the verification of public data sets and comparison with models of other algorithms, it is proved that the model of this study has a higher recognition accuracy than other recognition models, has good recognition effects on various abnormal driving behaviors, and the average recognition accuracy The rate is about 97%, which allows the identification of abnormal driving behaviors of intelligent network-connected vehicles, providing new ideas for ensuring traffic safety.