视觉预警与次任务对自动驾驶接管的影响研究
Research on the Impact of Visual Warning and NDRT on Autonomous Driving Takeover
朱浩宇1
作者信息
- 1. 淮北理工学院,机电与建筑工程学院 安徽 淮北 235000
- 折叠
摘要
L3 自动驾驶允许驾驶员从事与驾驶无关的次任务,次任务如何影响接管行为,预警信号在次任务干扰下是否仍然有效,类似问题研究甚少.为了探讨次任务下,预警信号在自动驾驶过程中对驾驶员接管行为的影响.基于驾驶仿真平台(UC-win/Road),设计了视觉次任务(4x4 箭头任务)下不同内容的视觉预警(简单视觉、行动视觉)的自动驾驶接管场景,进行 90 名被试的模拟驾驶试验.使用单、双因素方差分析,从驾驶行为、眼动行为2个接管绩效角度进行评析.研究结果表明:视觉次任务的加入,不会影响到行动视觉预警的优势效果;视觉预警相比无预警的手动能更好地引起驾驶员的注意,进而了解当前险情情况,且行动视觉预警的效果要优于简单视觉预警;其次,视觉预警会加重眼睛工作负荷.
Abstract
The influence of secondary tasks on takeover behavior and the effectiveness of early warning signals under their interference during L3 automatic driving have been rarely studied.To explore the impact of early warning signals on driver's takeover behavior in automatic driving scenarios,a simulated driving test was conducted with 90 subjects using the UC-win/Road driving simulation platform.Different visual subtasks(4x4 arrow task)were designed to examine the effects of visual early warnings(simple vision and action vision).Single-factor and two-factor analysis of variance were employed to evaluate both driving behavior and eye movement behavior during takeovers.The results indicate that adding visual subtasks does not diminish the superior effect of visual early warnings on driver takeover behavior.Specifically,action visual warning performs best,followed by simple visual warning.In terms of eye movement,visual warnings attract drivers'attention better than manual warnings or no warnings at all,enabling them to perceive current danger situations;moreover,mobile visual warnings are more effective than simple ones.Additionally,it is found that visual early warnings increase eye workload,with mobile visuals having a more pronounced effect.
关键词
自动驾驶/接管行为/次任务/视觉预警Key words
Automatic Driving/Takeover/Non-driving Related Task/Visual Warning引用本文复制引用
基金项目
2023年安徽省高校自然科学研究项目(2023AH052039)
出版年
2024
NETL
NSTL
万方数据