Research on Multimodal Takeover Interaction for Autonomous DrivingResearch on Multimodal Takeover Interaction for Autonomous Driving
Jiang Yanji1,2,3,Xie Rui2,Wang Xinrui 1,3,Zhang Yuqing2,Dong Hao1,3* 1Suzhou Automotive Research Institute,Tsinghua University;Suzhou,P.R.China 2Liaoning Technical University;Huludao,P.R.China 3OpenSafe Lab, Utcer (Suzhou) Electronics Technology Co.,Suzhou,P.R.China
Abstract—In the problem of switching control of L3-level autonomous driving, how to effectively remind the driver to take over the vehicle in an emergency is the key to ensure driving safety. In order to explore the takeover efficiency of different modal interaction modes, this paper designs six interaction modes-visual (V), auditory (A), tactile (T), visual-auditory (VA), visual-tactile (VT), and auditory-tactile (AT)-in an autonomous driving scenario, and conducts experimental testing and analyzing the takeover process, combining the driver's Response efficiency indexes analyze the changes of physiological signals with the takeover process, and construct a multi-criteria decision analysis model to evaluate the efficiency performance of different interaction modes. This study provides a scientific basis for optimizing the design of human-computer interaction for autonomous driving, and provides an important reference for improving the safety of takeover design in autonomous driving systems. Keywords-autonomous driving takeover; multimodal interaction; multicriteria decision analysis; multidimensional analysis method |