Driver's Personal Emotion Recognition for Intelligent Cockpit of New Energy Vehicles

Abstract

New energy vehicles pay attention to the development of intelligent driving technology, which requires high human-computer interaction technology. Therefore, this paper studies the emotion recognition of intelligent cockpit of new energy vehicles, so as to promote the emotion regulation of intelligent driving on drivers and assist drivers to complete driving behavior more efficiently and safely. In order to ensure that the angry driving data can be organically combined with the overall architecture of intelligent vehicles, the motion preview model based on integrated direction and speed control can make real-time decisions according to the vehicle motion state and current traffic information, and feedback the decision information to the vehicle control module. Based on this model, the driver's anger characteristics are considered. In addition, this paper proposes a multimodal driver emotion recognition model MDERNet based on facial expression and driving behavior. It filters and highlights the data of driving behavior modes through the temporal attention obtained from facial expression modes, so as to realize information fusion among multiple modes at the input information level. Finally, through the results of experimental research and analysis, we can see that the driver's personal emotion recognition proposed in this paper has a good effect.