A Study on Safety Issues of Human-Machine Transitions in Autonomous Vehicles

With the rapid development of artificial intelligence, sensor and computer, autonomous driving has become an important field and future trend of the automobile industry. However, it is still impossible to achieve the autonomous driving in all conditions in the short term, which may give rise to safety problems because of transitions between human and machine in the autonomous driving. This paper studied the definition and categorization of human-machine transitions, and proposed key issues concerning environment perception, personnel training and the transition mechanism by looking at the development of human-machine transitions in other industries. Also, this paper made a review of domestic and foreign literatures on these issues. The environment perception was found to be a key factor affecting human-machine transitions. There have been studies on factors that could influence the perception of drivers. However, the influencing mechanism remains unclear, preventing it from being applied in the design of human-machine transitions. Researchers have reached consensus on the significance of pre-use training on autonomous driving and compared the effects of various training programs. However, a comprehensive training system is not available yet, and there are no studies on the training criteria, preventing the accessibility of regulatory training in the short term. The design of human-machine transitions is still in the initial stage, consisting of a one-time reminder to the driver and waiting for taken over. Some researchers are working on the multi-step transition design and attempting to apply it in the real scenario, however its effects still need to be validated. Finally, the paper proposed suggestions on studies of human-machine transition based on domestic and foreign research.