In autonomous driving, the driver's responsibility for vehicle operation is reduced, which may lead them to engage in various non-driving related tasks. Understanding driver behavior during emergency takeover events is therefore crucial. Previous studies often used specific artificial tasks to distract drivers, neglecting the impact of realistic driving scenarios on interaction needs. This study developed a 3×4 lane-changing obstacle avoidance simulation experiment, incorporating cognitive load from non-driving tasks and warning system modes as variables. A total of 30 drivers participated in the trial, during which data were collected and analyzed. The results of the study showed that the Abstract Voice & Vibration TOR was the most effective at prompting drivers to take control of their vehicles, providing the best takeover times, lane change durations, and shortest crash times. While most participants favored the voice TORs, their takeover performance was relatively poor in the voice condition. In addition, there was a significant difference in takeover performance between low- and medium-load NDRTs, while there was no statistically significant difference in takeover performance between medium- and high-load NDRTs. These results emphasize the importance of considering the effects of auditory and haptic TORs when designing takeover requests and, at the same time, inform researchers and designers developing multimodal warning systems for self-driving cars.
The Impact of Non-Driving Related Task Types on Takeover Performance: A Simulated Driving Study Considering Multimodal Warning Systems
2024-10-25
1567981 byte
Conference paper
Electronic Resource
English
SAFE WARNING SYSTEM FOR AUTOMATIC DRIVING TAKEOVER AND SAFE WARNING METHOD THEREOF
European Patent Office | 2019
|Safe warning system for automatic driving takeover and safe warning method thereof
European Patent Office | 2020
|