Design Guidelines for Voice Driving Assistant Agents Based on Driver-Passenger Interactions

Drivers must maintain awareness of traffic conditions to prevent accidents, however it is difficult to perceive all relevant hazards in real time. Passengers often mitigate this limitation by warning drivers about blind spots or unexpected events. Building on this insight, this project proposes design guidelines for a voice driving assistant agent that emulates a helpful passenger. We conducted an observational study of driver–passenger interactions across varied road and traffic conditions to characterize how passengers support driving. The analysis shows that passengers provide adaptive assistance (e.g., situational awareness, traffic rules, and route guidance) by considering the driver's characteristics (e.g., driving ability, psychological state, physical condition, driving habits). We synthesize these findings into design guidelines for in-vehicle voice driving assistant agents for enhancing situational awareness.