Optimized HMI Strategies for Collaborative Driving Interfaces in L3+Autonomous Vehicles

Amidst the pivotal transition of autonomous vehicles from Level 2 to Level 3, a fundamental paradigm shift occurs, moving from a shared human-machine control framework to a complete transfer of driving responsibilities. In this evolving context, the automated system assumes a primary role in executing driving tasks, while the driver transitions to a supervisory role, overseeing system behavior and intervening only when deemed necessary. This transition necessitates addressing complex challenges encompassing safety, liability, workload distribution, alongside fostering human-machine trust and enhancing system interpretability. To address these multifaceted challenges, we introduce a comprehensive human-machine shared driving interaction design strategy, referred to as the TSE design method. This strategy encapsulates the entire driving journey, commencing with pre-journey training and preparation, encompassing safe driving practices during the journey, and culminating in post-journey performance evaluation. Our primary emphasis lies in the intricate analysis of human-machine safety interaction strategies during the actual driving phase. Based on the current system status and road conditions, the automated system dynamically apportions driving tasks to the driver, considering factors such as the driver's workload, engagement level, and the complexity of take-over tasks. Furthermore, we delve into the intersectional impact of various factors, including the driver's mental model, attention allocation, situational awareness, human-machine trust, system transparency, and interpretive interfaces, on the overall performance of the human-machine shared driving system. To demonstrate the practical application of our design principles, we present concrete design exemplars derived from driver training sessions and take-over interaction design. These instances illustrate the effectiveness of our TSE design method in enhancing the safety, trust, and interpretability of the human-machine shared driving system. This paper presents a comprehensive and systematic methodology for the design of L3+ autonomous driving human-machine interactions, thereby offering profound insights and cutting-edge solutions to automotive manufacturers in their pursuit of optimizing human-machine interaction design endeavors.