The Rise, Fall, and Repair of Trust for Automated Driving Systems

The purpose of this study was to investigate how human driver\u27s trust in the automated driving system is built over time and affected by automation failure. The study expanded trust development over time by measuring trust after a practice demonstration ofthe system capabilities and after each of seven unique, sequential drives. The automation performed perfectly on six of the seven drives but made one of three different responses to a critical hazard event in the fourth drive. Depending on the error-type condition, the automation either perfectly avoided the hazard (no error), issued a takeover request (TOR), or failed to notice the hazard (failure). In contrast to the typically used pre/post trust-difference assessment that does not show a trajectory of growth or decline patterns, the current design allowed for evaluation of the growth, decline, and repair of trust