可控性
控制(管理)
计算机科学
工程类
系统工程
可靠性工程
人工智能
控制工程
生化工程
数据挖掘
数学
应用数学
作者
Frederik Naujoks,Katharina Wiedemann,Nadja Schömig,Oliver Jarosch,Christian Gold
出处
期刊:MethodsX
[Elsevier BV]
日期:2018-01-01
卷期号:5: 579-592
被引量:41
标识
DOI:10.1016/j.mex.2018.05.007
摘要
Up to a level of full vehicle automation, drivers will have to be available as a fallback level and take back manual control of the vehicle in case of system limits or failures. Before introducing automated vehicles to the consumer market, the controllability of these control transitions has to be demonstrated. This paper presents a novel procedure for an expert-based controllability assessment of control transitions from automated to manual driving. A standardized rating scheme is developed that allows trained raters to integrate different aspects of driving performance during control transitions (e.g., quality of lateral and longitudinal control, adequateness of signalling to other road users, etc.) into one global controllability measure based on video material of the driving situation. The method is adapted from an existing assessment procedure that has been successfully applied to assess the criticality of driving situations in manual driving conditions (e.g., assessment of substance-induced impairments, assessment of fitness-to-drive of novice drivers). This paper presents the rating procedure, including instructions of how to code relevant qualities of the drivers' performance with accompanying video-demonstrations, and material used for rater training. •A rating procedure for an expert-based controllability assessment of control transitions from automated to manual driving based on observation of video material was adapted from an existing method used in studies on manual driving.•The advantage of this method consists in an integration of different dimensions of driving performance (e.g., operational and tactical driving behaviour, criticality of the situation) into one global controllability measure.•The method allows an assessment and comparison of diverse take-over scenarios, detached from driver performance variables.•The accompanying video-based training material allows reproducible and reliable execution of the rating procedure.
科研通智能强力驱动
Strongly Powered by AbleSci AI