Effects of Control Transition Strategies and Human-Machine Interface Designs on Driver Performance in Automated Driving Systems

Abstract

This study addresses an issue arising during requests to intervene, usually focused on the direct shift of control from the system to an unprepared driver. We explore an approach based on the concept of adaptive automation. The approach is utilized to establish a strategy that the automated driving system can modulate between different driving states (Level 0 (L0) to Level 3 (L3) driving automation) based on traffic situations. Furthermore, we aim to design Human-Machine Interface (HMI) instructions that align with two design perspectives: system-centered (Implicit) and human-centered (Explicit). These HMI instructions critical to the strategy implemented are intended to sustain awareness on driver roles and decrease inappropriate driver responses during transitions. In the experiment, we utilized two approaches (adaptive control transition strategy and fixed control transition strategy) and two HMI designs. The experiment with a driving simulator was conducted with 60 drivers. The research questions explored include the influence of HMI design on drivers’ awareness during shifts from L3 to L2/L1. Also, we examine how different combinations of a strategy and HMI design affect reaction times from L3 to L0 scenarios. Compared to Implicit-HMI, the use of Explicit-HMI did not lead to inappropriate actions associated with their driving roles. This indicates that Explicit-HMI facilitates accurate response to state changes during the implementation of the adaptive approach. In terms of the comparison of the combinations, it is found that the adaptive approach provides initial guidance for drivers to monitor the environment, thereby providing earlier reaction time for the transition to manual.