人机共驾智能汽车的控制权切换与安全性综述

根据智能汽车技术发展特点和趋势提出了人机共驾的概念; 从切换的发起者、强制性与计划性三方面论述了人机共驾智能汽车控制权切换的分类方法, 分析了广义和狭义2种分类的特点和应用范围; 从驾驶人的认知、驾驶负荷、反应力等方面剖析了人机共驾中人因的特性及其对控制权切换安全性的影响, 总结了控制权切换的试验研究方法和人机交互形式, 指出了控制权切换安全性研究存在的问题和未来发展方向。分析结果表明: 人机共驾智能汽车的应用范围是L2~L3级自动驾驶, 特点是人与系统彼此协同完成动态的驾驶任务; 由系统主动发起、驾驶人被动接管的控制权切换情形与安全性更被业内关注; 驾驶人能有效地对当前驾驶状态进行认知和评估, 进而接管车辆操作, 并最终规避风险, 是保证控制权切换安全性的关键; 人因是影响控制权安全平稳切换的重要因素, 主要表现为认知水平偏低, 切换前后驾驶负荷阶跃式突变, 次任务的影响机理不明确, 反应力随切换场景的不同而差异显著等; 该领域的主要研究还包括接管绩效的评价, 切换时机与人机交互方式的优化以及试验手段的提升等。 

Review of control switch and safety of human-computer driving intelligent vehicle

On the basis of the features and tendencies of intelligent vehicle technology, the concept of human-computer driving was proposed.The taxonomy methods of the control switch of intelligent vehicles were analyzed in terms of the initiator, urgency, and schedule.The features and applications of two taxonomy methods were discussed in a generalized and narrow sense.The characteristics of human factors in the field of human-computer driving and its influence on the safety of control switch were analyzed in terms of the driver awareness, workload, and response.The experimental methods and human-machine interaction forms of control switch weresummarized, and the problems and further developments of the safety were pointed.Analysis result indicates that the application scope of human-computer driving intelligent vehicle is L2-L3 automated driving, and the characteristic is the cooperation of human and computer to complete the dynamic driving tasks.The safety of the control switch situations, which is initiated by the system and passively taken over by the driver, has been given special attention.The key to ensure the safety of control switch is that the driver can effectively recognize and evaluate the current driving state, take over the vehicle operation, and finally avoid the potential hazard.Human factors have a severe impact on the safety and stability of the control switch, mainly manifested as the loss of situation awareness, sudden change of workload before and after handover, uncertainty of the influence of secondary tasks, and significant differences of driver response in different scenarios.The main research in this field also include the evaluation of taking over performance, optimization of switching timing and human-machine interaction, and the improvement of experimental methods.