面向人机共驾车辆的驾驶人风险感知研究综述  

A Reviewon Driver’s Perception of Risk Associated with Autonomous Driving Under Human-computer Shared Control

作　　者：冯忠祥 李靖宇 张卫华 尤志栋 FENG Zhongxiang;LI Jingyu;ZHANGWeihuang;YOU Zhidong(School of Transportation,Southeast University,Nanjing 210096,China;College of Civil Engineering,Hefei University of Technology,Hefei 230009,China;School of Automobile and Traffic Engineering,Hefei University of Technology,Hefei 230009,China;Traffic Management Research Institute of the Ministry of Public Security,Wuxi 214000,Jiangsu,China)

机构地区：[1]东南大学交通学院,南京210096 [2]合肥工业大学土木与水利工程学院,合肥230009 [3]合肥工业大学汽车与交通工程学院,合肥230009 [4]公安部交通管理科学研究所,江苏无锡214000

出　　处：《交通信息与安全》2022年第2期1-10,共10页Journal of Transport Information and Safety

基　　金：国家自然科学基金项目(71971073);中国博士后科学基金面上项目(2019M661694);公安部道路交通安全重点实验室开放基金项目(2021ZDSYSKFKT01);认知智能国家重点实验室开放基金计划项目(CIOS-2020SC04)资助。

摘　　要：面向人机共驾车辆的驾驶人风险感知是接管时正确应激反应和操作的前提,是交通安全领域的研究重点。分析了人机共驾车辆驾驶人风险感知概念及其特性;从驾驶人特性、自动驾驶系统、驾驶情景这3个方面分析了人机共驾车辆驾驶人风险感知的影响因素;从驾驶行为表现、接管绩效和主观评价这3个方面对人机共驾车辆驾驶人风险感知衡量方法进行归纳总结;梳理归纳了基于驾驶人培训、辅助设备调节的风险感知能力提升方法。结果表明:相比于手动驾驶,人机共驾车辆驾驶人风险感知能力较低,且是多因素耦合作用下的结果;现有风险感知能力评价方法各有弊端,缺少可广泛应用的普适性量化方法;对驾驶人状态进行动态监测和调节是保障人机共驾车辆安全应用的前提。基于现有研究中存在的问题,指出了人机共驾车辆驾驶人风险感知未来研究方向,主要包括多因素耦合情况下的风险感知研究、风险感知能力量化模型构建、风险感知能力安全阈值研究、风险感知能力动态监测与稳态保持方法研究。Timely perception to risk associated with autonomous driving under human-computer shared control is the premise of the correct stress response and operation of drivers,and it is the focus of road safety research. The characteristics of risk perception for drivers of human-computer shared control are analyzed. Influencing factors are analyzed from three aspects:driver’s characteristics,automatic driving system,and driving scenario. Besides,evaluation methods are analyzed and summarized from the following three aspects:driving behavior,take-over performance,and subjective evaluation. Moreover,improvement methods for increasing the ability of risk perception through driver training and auxiliary equipment are summarized. Study results show that compared with manual driving vehicles,the capability of drivers’ risk perception to human-computer interaction during the operation of autonomous vehicles is lower,which results from the interactions of multiple factors. The existing methods for evaluating the capability of driver’s risk perception have their own advantages and disadvantages,and there is no universally applicable method that can be widely used. Dynamic monitoring and adjustment of driver’s state is the safety prerequisite of autonomous driving under human-computer shared control. Based on the issues identified from the existing studies,it can be concluded that future studies should address the following:risk perception under the interaction of multiple factors,quantitative modeling of the capability of driver’s risk perception,dynamic monitoring,and steady-state maintenance methods for driver’s risk perception.

关 键 词：交通安全 自动驾驶 人机共驾 风险感知 驾驶行为 

分 类 号：X951[环境科学与工程—安全科学]