Trust in Automated Vehicle: A Meta-Analysis

AbstractTrust in automation has gained much attention in both industry and academia. More and more studies and evidence prove its importance in new technology acceptance and efficient human-automation cooperation. As one of the most eye-catching and complicated intelligent systems ever made by humans, automated vehicles (AVs) will change people’s daily lives and promote next-generation transportation. Trust in AVs then becomes a critical research topic towards the efficient and safe implementation and utilization of such systems. With more studies published in the AV research area focusing on different factors toward trust, there lacks a systematic summary of the state-of-the-art findings to build the current frontier and guide future research. This study conducts a meta-analysis on more than fifty antecedents identified from more than two hundred related publications in recent years. We firstly classify trust factors collected from these studies into three main categories: human-related, AV-related, and environment-related. Human-related factors include ability-based factors and characteristics, and AV-related factors include performance-based and attribute-based ones. The classification process enables us to generalize the factors found in the literature and synthesize and analyze corresponding effects. Results show that human-related factors are significantly affecting trust with the highest correlation score, and human characteristics are the most influential factors. AV-related factors are also significant towards trust, with attribute-based factors being more influential than performance-based factors. Environmental factors are less studied in the publications. These findings could guide AVs’ development, design of user training, and future research directions of trust in AVs.KeywordTrust in automationAutonomous drivingMeta-analysis