Autonomous vehicles (AVs) hold the potential to revolutionize transportation, offering benefits such as reduced crashes, traffic congestion, and greenhouse gas emissions. However, significant uncertainties surround their future adoption. This study addresses the need to understand the barriers hindering widespread AV adoption, focusing on the complex, interconnected nature of these barriers, moving beyond pairwise relationships to explore causal influences. The research questions are: (a) What are the key barriers to AV adoption? (b) How do these barriers causally influence each other? (c) How can these causal influences be depicted and analyzed?

Existing research on AVs has focused on opportunities and challenges, consumer willingness to pay, and system-level impacts. Most studies have relied on stated preference surveys and econometric analyses, often examining pairwise relationships between barriers. This study extends previous research by considering the multifaceted relationships between multiple barriers simultaneously and exploring causal, rather than merely associative, relationships. A review of existing literature identified ten distinct barriers to AV adoption, including reduced security and privacy, social inequity, obscurity in accountability, lack of customer acceptance, potential loss of employment, inadequate infrastructure, lack of standards, absence of regulation and certification, manufacturing cost, and induced travel.

The study employs a five-stage methodology. First, ten barriers to AV adoption are identified through a literature review and expert consultations. Second, a survey of eighteen experts from academia and industry gathers data on the interrelationships between these barriers using a linguistic scale. Third, Grey-DEMATEL, a multicriteria decision-making technique that accounts for uncertainty, is applied to the expert data to rank the barriers and identify cause-and-effect relationships. Fourth, a sensitivity analysis assesses the robustness of the results using different expert weighting schemes. Fifth, the findings are presented using a causal loop diagram (CLD), a systems thinking approach, to understand the dynamic interplay of barriers. The study uses grey system theory in conjunction with DEMATEL because grey theory generates satisfactory results when dealing with limited, uncertain, or incomplete data. This enhancement enhances the accuracy of human judgment when integrated into the decision-making process. The US is chosen as the geographical context due to its leading role in AV innovation and its challenges in AV adoption.

The Grey-DEMATEL analysis reveals that lack of customer acceptance (LCA) is the most prominent barrier (highest R+C score), indicating high influence on and susceptibility to other barriers. It is also identified as the greatest net effect barrier (lowest R-C score), meaning that other factors influence it the most. Manufacturing cost (MNC) is identified as the most influential cause barrier (highest R-C score). The causal loop diagram (CLD) illustrates the complex interplay among barriers, revealing ten feedback loops. Six loops directly involve LCA, indicating its central role. The CLD suggests that addressing the lack of standards (LOS) and the absence of regulations and certifications (ARC), which are also prominent, could indirectly improve customer acceptance.

The findings highlight the interconnectedness of barriers to AV adoption, emphasizing the need for holistic policy interventions rather than addressing individual barriers in isolation. The prominence of LCA underscores the importance of building public trust and addressing safety concerns. The study's integration of Grey-DEMATEL and systems thinking offers a novel approach to analyzing complex systems of barriers, allowing policymakers and stakeholders to prioritize efforts based on causal relationships, not just pairwise correlations. The sensitivity analysis confirms the robustness of the key findings, despite variations in expert weighting. The study's focus on causal relationships provides a more nuanced understanding than previous research primarily focusing on pairwise associations.

This study provides valuable insights into the key barriers to AV adoption, particularly emphasizing the central role of lack of customer acceptance and the importance of addressing the lack of standards and regulations. The methodology integrates Grey-DEMATEL and systems thinking, offering a robust framework for analyzing complex systems of barriers. Future research could explore community acceptance, integrate the DEMATEL framework with econometric modeling, and expand the geographical scope of the study.

The study relies on the opinions of eighteen experts in the US, limiting generalizability. While the sensitivity analysis supports the robustness of findings, a larger, more diverse expert panel could yield more comprehensive results. Future research should investigate other relevant barriers such as community acceptance and expand the geographical scope to consider diverse contexts.