Trust evaluation of automated vehicles: A systematic review

The rapid development of automated vehicles offers the promising development of driver-vehicle interaction and cooperation. Trust is an important concept to consider for the future implementation of autonomous driving. An inappropriate level of trust can lead drivers to under-trust and reject the system's potential benefits or allow drivers to over-trust and abuse it. Therefore, autonomous vehicles need an appropriate level of trust for drivers to experience the full benefits of autonomous driving. This paper reports a systematic review of the literature to analyse the critical role of trust and also discusses various methods of evaluating the trust between drivers and automated vehicles to promote the use of autonomous driving on the ground.The review surveyed the trust in automated vehicles and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. First, the importance of trust in increasing the acceptance of autonomous driving is investigated. Second, the factors influencing drivers' trust in autonomous driving are grouped and presented. The analysis focuses on individual driver characteristics, automated vehicles and the driving environment, such as driver preference, driving automation system and driving scenarios. Finally, the methodologies to measure trust in autonomous driving are reviewed and analysed. The key measurement indicators include questionnaires, physiological signals such as eye gaze, head and body postures, etc. and psychological signals such as electroencephalogram (EEG).This study is expected to summarise the factors that influence trust and to find reliable and replicable methods to measure trust. The results show that the influence of different factors on trust varies considerably. Currently, questionnaires are the most commonly used subjective measurement method, while psychophysiological measures are a promising objective complement and attract increasing investigations.