Improving Human-Autonomous Vehicle Interaction in Complex Systems

TL;DR

This dissertation reframes human-AV interaction as a complex, context-sensitive system and tests how communications should adapt to rider goals, contexts, and traits. Across three empirical studies, it demonstrates that task- and modality-appropriate multimodal explanations can support learning, that explanation errors reduce trust depending on driving context, and that personal traits strongly predict trust, with risk-benefit perceptions largely driving acceptance. It provides design guidelines for transparent, adaptable, and personalized AV explanations, supported by robust ML analyses and SHAP interpretability. The work advances a holistic framework for human-AI collaboration in safety-critical domains and offers actionable insights for designers, researchers, and policymakers to improve the real-world adoption of autonomous vehicles.

Abstract

Unresolved questions about how autonomous vehicles (AVs) should meet the informational needs of riders hinder real-world adoption. Complicating our ability to satisfy rider needs is that different people, goals, and driving contexts have different criteria for what constitutes interaction success. Unfortunately, most human-AV research and design today treats all people and situations uniformly. It is crucial to understand how an AV should communicate to meet rider needs, and how communications should change when the human-AV complex system changes. I argue that understanding the relationships between different aspects of the human-AV system can help us build improved and adaptable AV communications. I support this argument using three empirical studies. First, I identify optimal communication strategies that enhance driving performance, confidence, and trust for learning in extreme driving environments. Findings highlight the need for task-sensitive, modality-appropriate communications tuned to learner cognitive limits and goals. Next, I highlight the consequences of deploying faulty communication systems and demonstrate the need for context-sensitive communications. Third, I use machine learning (ML) to illuminate personal factors predicting trust in AVs, emphasizing the importance of tailoring designs to individual traits and concerns. Together, this dissertation supports the necessity of transparent, adaptable, and personalized AV systems that cater to individual needs, goals, and contextual demands. By considering the complex system within which human-AV interactions occur, we can deliver valuable insights for designers, researchers, and policymakers. This dissertation also provides a concrete domain to study theories of human-machine joint action and situational awareness, and can be used to guide future human-AI interaction research. [shortened for arxiv]

Figures (19)

• Figure 1: Overview of the Human-AV system. Action goals, the driving context, and different traits (human and AV traits) are fundamental parameters determining what actions -- communicative and not -- need to be jointly taken by the team. Actions build, and are iteratively impacted by, the different types of situational awareness held by the system as the team works towards goal success.
• Figure 2: Example Success function for the goal of Safe Transportation.
• Figure 3: Outline of the premises of this dissertation. This dissertation is emergent from the intersection of two problems: (1) failure to support goals and (2) complex systems change success criteria needed to meet goals. By bridging these two problem areas, we can improve human-AV interaction. Each chapter in this dissertation looks at different areas of this problem-solution puzzle.
• Figure 4: Full-motion driving simulator.
• Figure 5: Visual 'what' racing line projection. The green racing line projected on the track is an example of a visual ‘what’ explanation seen by Group 4.