Using a Virtual Agent to help Distracted Drivers notice Hazards in Self-driving Cars

What is it about?

As self‑driving vehicles becomes more advanced, drivers are allowed to take their hands off the wheel and focus on other tasks. However, they still need to be ready to take control if something dangerous happens on the road. Most cars rely on visual warnings or audio warnings to alert drivers, but these do not always help the driver to see where the danger is, especially if they are distracted. Virtual agents, such as those used in navigation or voice‑assistant systems, offer another way to provide information with drivers. In our research, we have shown that drivers are able to use visual cues and social cues from a holographic virtual agent to help them predict danger on the road, even when they were distracted by another task and at the same level as when they were fully focused on watching the road.

Featured Image

Photo by Erik Mclean on Unsplash

Photo by Erik Mclean on Unsplash

Why is it important?

These findings show that drivers who were distracted by a video game were able to use attentional cues towards to recognise imminent danger on the road to the same level as when fully focused on the road. This suggests that is possible to design interface that allow drivers in self-driving cars to engage with non-drivings tasks like using their phone or playing a game while also being kept in the loop about the road. Additionally, these findings suggest that holographic virtual agents displaying social cues (such as looking towards hazardous road events) could also be a useful tool on top of visual cues in helping to keep drivers aware of the road and improve safety in automated vehicles.

Perspectives