Research On Driver Emotion For Smart Cockpit: Behavior Impact,Recognition And Regulation Method

With the development of the level of intelligence and connectivity,the intelligent and connected vehicle is transforming into multiple intelligent mobile terminals that carry rich functions and services.This development will expand and deepen the scope of human-computer interaction between humans and vehicles,thereby creating a new and more natural human-vehicle relationship in the smart cockpit.The continuous development of the smart cockpit will also help solve the driver's emotional out-of-control and improve driving safety and comfort.Automotive smart cockpit is an tntelligent service system that refers to the needs of users in the application scenarios,equipped with intelligent and connected in-vehicle products or technologies,with insight,understanding,and meeting user needs,and can achieve safe,efficient,comfortable and pleasant HMI experience.The research of driver emotions in the "emotion-aware interaction" smart cockpit is one of the effective ways to achieve a safe,efficient,comfortable and pleasant driving experience.There are four common problems in the current research on driver emotion: First,the previous research lacks a comprehensive discussion of driver emotion mechanism and the construction of a complete driver emotion research framework.Second,the difference in the influence of different emotions on driver behavior is unclear.The third is the insufficient research on the non-contact driver's multiple emotion recognition methods in dynamic scenarios.In addition,the previous research lacks relevant data sets for the research on driver emotion recognition.Fourth,there is insufficient research on the mechanism of driver emotion regulation.In response to the above problems,this paper studies the mechanism and framework of driver emotions;the influence of driver emotions on driving behavior;driver emotion recognition and driver emotion data sets based on facial expressions and driving behavior;driver's emotion regulation based on visual elements and their application prospects.The mechanism and framework of driver emotion research are the basis of driver emotion research.This article first elaborates the basic knowledge of emotion research,the mechanism of emotions affecting driving,and emotions from four aspects: the development,generation process and classification of emotions,the influence process of emotions on driving,the multimodal expression of emotions,and the adjustment process of emotions.The recognition mechanism of emotion,the mechanism of emotion regulation strategy.On the basis of clarifying the research mechanism of driver emotions,this paper constructs a framework for the study of driver emotions in the "emotion-aware interaction" smart cockpit,which includes three layers: driver,smart car and network.The driving behavior caused by the driver's emotion is one of the important factors affecting driving safety.In addition,the analysis results of the influence of driver's emotions on driving behavior will provide references for designing driver's different emotion recognition schemes and adjustment strategies,thereby enhancing the driving experience.Starting from the discrete emotion model and the dimensional emotion model,this chapter mainly studies the impact of Chinese drivers on driving behavior under different emotions.This paper first constructed a library of driver emotion-inducing materials,and then implemented a variety of emotional driving experiments for drivers.Finally,through data analysis,the results of the impact of discrete emotions on driving behavior and the results of dimensional emotions on driving behavior were obtained.Driver emotion recognition is an important research content of "emotion-aware interaction" smart cockpit.This paper first proposes a multi-modal driver emotion recognition model MDERNet based on facial expressions and driving behavior.Then we implemented the data collection experiment of facial expression and driving behavior data under the driver's emotional driving,and finally used the collected data to train the model to verify the effectiveness of the MDERNet model.At the same time,this chapter also analyzed the model on other public data sets.Universality and advancement have been verified.In addition,this paper also established a multi-modal data set of driver emotions in a simulator environment,a data set of driver emotional facial expressions and road scenes in a real vehicle environment.These data sets are helpful in the field of driver emotion recognition.the study.Regulating the driver's emotions is an important means to realize the "emotion-aware interaction" smart cockpit.Aiming at the driver's anger,which is the main reason that affects driving safety,this paper studies the anger adjustment quality of different parameters of different visual elements in the driving process,from the perspective of visual design parameters,from the perspective of color(cold/warm)and form of expression(The three dimensions of flatness/simulation)and adjustment methods(active intervention/state feedback)are used to study the regulation of driver's anger emotion.This article first designs a driver's anger adjustment program based on the above parameters,then implements the driver's emotion adjustment experiment,and then discusses the driver's emotion adjustment quality in the three dimensions of color,expression and adjustment method.Finally,the application prospects of driver emotion research are discussed.