定性与定量分析跟驰行驶中汽车驾驶员情感-行为交互作用机理

Understanding the interaction between drivers’ emotions and driving behavior is essential for developing anthropomorphic autonomous vehicles and achieving intelligent human-machine interaction in connected vehicles. Previous research has focused on the impact of emotions on driving behavior, but the effects of driving behavior on driver emotion remain largely unexplored. From a new perspective of driving behavior affecting drivers’ emotion, this proposal will deeply analyze the interaction between drivers’ emotions and driving behavior during car-following from the perspective of cognition under a microscope. The study will conduct the real driving experiments and driving simulations to collect the human-vehicle-environment data, and apply the qualitative methods of think aloud、survey questionnaire、individual interview and focus group interview to collect drivers’ emotional and cognitive data. Grounded theory method will be used to investigate the mechanism of the interaction between drivers’ emotions and driving behavior. Structural Equation Model will be used to analyze the interaction effects of drivers’ emotions and behavior. Hidden Markov Model will be used to determine the transformation mechanism of driving behavior triggered by drivers’ emotions. By integrating the quantitative and qualitative results, an interaction model between emotions and behavior will be established based on drivers’ cognitive paths and sequences on processing traffic information during driving. The findings of this study can provide strong support for the development of personalized driving warning system and intelligent human-machine interaction.

研究驾驶过程中的情感特点和行为模式，正确认识驾驶情感与行为之间的影响作用，是实现拟人化无人驾驶和智能网联环境下人车和谐交互的重要基础。目前研究大多关注在驾驶情感对于行为的影响，而忽视了驾驶行为影响情感的研究。本项目拟从驾驶行为影响情感这一新视角出发，深度剖析在跟驰行驶中驾驶情感与行为在认知范畴内的微观交互作用机理，并定量分析驾驶情感与行为的影响作用效应。本研究通过模拟驾驶和实车道路实验，运用think aloud、问卷、个人深度访谈和焦点小组访谈等定性方法收集内隐驾驶员情感认知数据。使用扎根理论探究驾驶情感与行为的交互作用机理；运用结构方程模型剖析驾驶情感与行为的作用效应；采用隐马尔科夫模型推理情感演变激发的驾驶行为转移规律。在定性和定量分析的基础上，根据驾驶员认知路径和序列构建驾驶情感与行为交互作用模型。本研究阐明的作用机理和构建的交互模型，为车辆预警系统和人机交互的发展提供理论依据。

研究驾驶过程中的情感特征和行为模式，正确认识驾驶情感与行为之间的影响作用，是实现拟人化无人驾驶和智能网联环境下人车和谐交互的重要基础。本项目从情感与行为交互这一新视角出发，在驾驶人情感-行为交互作用机理和模型构建等方面取得了一系列成果。首先，运用结构方程模型构建驾驶行为状态下的情感效应模型，和驾驶情感状态下的行为效应模型，定量剖析驾驶情感与行为之间的作用大小，并揭示其作用实现路径和深层动因。其次，采用隐马尔可夫模型，推演驾驶情感-行为双重状态转移链，解析驾驶情感演变激发下的驾驶行为转移规律，预测驾驶员行为意图的动态演变。再次，通过深度挖掘驾驶人的生理特征指标，构建了基于认知路径和序列的驾驶情感与行为交互作用模型，从能量层面明晰了其交互作用机理。本项目所取得的研究成果，拓展了驾驶情感研究的思路，为智能汽车人机交互的开发提供理论依据。