基于动态聚类算法的情绪对模拟驾驶仿真视觉关注度影响研究

A Study on the Effect of Emotion on Visual Attention in Driving Simulation Based on Dynamic Clustering Algorithm

摘要: 为防范并减缓因驾驶员情绪波动所导致的交通事故，本研究探讨了模拟仿真驾驶过程中情绪对驾驶员视觉关注度的影响。首先诱发被试人员产生正常情绪与愤怒情绪。其次基于眼动实验分析两组被试人员的视觉注意关注状态，得到了两组被试人员的眼动轨迹图、热点图、眼跳等眼动指标进行分析情绪对视觉关注的影响，之后利用动态聚类算法对注视点进行聚类，呈现出被试人员不同的感兴趣的区域，分析被试人员对车外信息处理情况对比。最后结果表明：愤怒情绪下的被试人员处理车外信息较慢，视野活动轨迹较窄，反应速度较慢。本研究有助于更深入地探究愤怒情绪与道路行驶安全之间的联系，并为开发有效的交通安全策略提供科学依据。

Abstract: In order to prevent and mitigate traffic accidents caused by drivers’ emotional fluctuations, this study investigated the effects of emotions on drivers’ visual attention during simulated driving. Firstly, the subjects were induced to produce normal emotions and angry emotions. Secondly, based on the eye movement experiment to analyze the visual attention focus state of the two groups of subjects, the eye movement trajectory map, hotspot map, eye hopping and other eye movement in-dexes of the two groups of subjects were obtained to analyze the effect of emotion on visual atten-tion, and after that, a dynamic clustering algorithm was utilized to cluster the focus points, pre-senting different regions of interest of the subjects, and to analyze the comparison of the subjects’ processing of information outside of the vehicle. The final results show that the subjects under the emotion of anger process out-of-vehicle information slower, have narrower trajectories of visual field activity, and have slower reaction speeds. This study helps to explore the link between anger mood and road driving safety more deeply, and provides a scientific basis for developing effective traffic safety strategies.