Research On The Influence Of External Human-machine Interface Of Automated Vehicles On Driving Behavior Based On Unprotected Left Turn

There will be a period of penetration for automated vehicles after their implementation in actual traffic,and will be a mixed manual-automated vehicle traffic flow for a long time during this period.In this mixed traffic flow,the social interaction between drivers and automated vehicles would be absent,which affects the safety and efficiency of vehicles’ interaction.Unprotected left turn is one of the vehicle interaction scenarios with high accident rates and complexity at intersections.In such complex interaction scenarios,which rely on social interaction and empirical judgment between drivers,the lack of information communication between autonomous vehicles and drivers is one of the urgent problems that need to be solved in the development of autonomous driving technology.Research shows that external humanmachine interface(e HMI)can effectively compensate for the lack of communication between the driver and the autonomous vehicle during the interaction process.Based on the above background,the purpose of this study is to investigate the effective communication between non-autonomous vehicle drivers and autonomous vehicles in complex interaction scenarios through e HMI.In this thesis,unprotected left-turn interaction process in real roads were analyzed firstly.The unprotected left-turn interaction data between vehicles were collected through field observations,and the observed data were compiled and analyzed to explore the influencing factors of right-of-way communication for unprotected left-turn;then,an unprotected left turn driving scenario at the intersection was built in the virtual platform,and the unprotected left turn influencing factors in the actual road were introduced as variables to explore the decisionmaking and behavioral characteristics of conservative and aggressive drivers on the self-driving car under different levels of variables;Finally,the e HMI information was added to the selfdriving car in the process of unprotected left turn interaction,and the scenario with the lowest driver interaction efficiency in the self-driving car interaction experiment was selected to test the influence of different e HMI presentation contents and text colors on the driver’s decision making behavior through driving simulation experiments to explore the effective presentation of e HMI information in complex scenarios.The results showed that 1)The main influencing factors of communication results between manual-driving car drivers for unprotected left-turn right-of-way in actual roads are the steering angle of left-turning vehicles and the complexity of the environment during the interaction.2)There are stage characteristics of manual-driving car drivers’ decision judgments during the current manual-autonomous car unprotected left-turn interaction.At the early stage of interaction,the driving decisions of manual-driving car drivers are related to driving experience;in the middle stage of interaction,the decisions of manual-driving car drivers are influenced by the yielding state of the autonomous car.3)Drivers with different driving styles tend to be consistent in their behavioral decisions when facing the autonomous car;during the unprotected left turn interaction,drivers maintain a skeptical and cautious attitude toward the behavior of the autonomous car.4)The e HMI information presentation facilitates the interaction performance between drivers and autonomous vehicles in unprotected left-turn situations.e HMI that presented explicit information had better performance and shorter decision time for drivers than e HMI that presented implicit information;the extension of the text color information of the traffic rule color system confused drivers’ understanding of the text content to some extent;some drivers paid less attention to the text color information.This paper suggests that automated vehicles using e HMI for social interaction with nonautonomous vehicles in complex interaction scenarios,and using explicit information to communicate with non-autonomous vehicles when designing the external display interface for autonomous vehicles,clearly indicating the autonomous vehicle’s own intentions and decision information;it is also suggested that reducing the influence of text color on the receiver,reducing the use of color systems with additional information,using neutral color systems to present information content,and avoiding information confusion caused by text color.