Get in touch

shawn's email

Developing in-car emotion detection & AI driver assistance

What if your car could autonomously detect driver and passenger mood or stress levels, in addition to road conditions? How then might it interact with drivers and passengers to improve the driving experience and safety in an emotionally intelligent way? What might a vehicle with such a dynamic personality look, sound, and feel like?

An illustration of a woman driving a car during a storm with a male passenger, while the dashboard bathes them in green light.

For the better part of a century, the basic human experience of driving a car has remained relatively unchanged — people must be alert, adhere to the rules of the road, and drive defensively. With the advent of a smartphone in everyone’s pocket (and now, one on our dash), there hasn’t been a more dangerous time to operate a vehicle. No matter how skilled the driver, or how many safety features a vehicle has, distracted driving can have fatal consequences.

Humanistic partnered with one of the large global automotive manufacturers and a massive technology company to explore how sensors in future vehicle cabins might be able to cut down on distracted driving, as well as provide timely guidance in stressful driving conditions. 

Through our exploration, it became clear that what was needed was some sort of intelligent (and emotionally affective) in-car agent — essentially a car with a dynamic personality like KITT from the hit 1980’s series Knight Rider. Such an in-car system would need to sense the situational context of the road, estimate the physical and mental states of the passengers and driver, and determine an appropriate emotional response. Many times this meant proactively interacting with the user to warn them or provide guidance, without a formal invoking by the occupants. 

a redacted flow diagram deliverable from the project
human affected research example
redacted flow diagram
redacted user journey and flow diagram
a point cloud graph demonstrating human affect
a redacted affective computing flow diagram
project notes
project notes
project notes
a human affect journey diagram
human affect diagram
research insights
No items found.

The result of this work has spurred a number of major follow-on projects that begin the long process of commercializing these new technologies — drawing on the fields of social psychology, emotional cognition, conversational frameworks, and the science of passively detecting human intent.


More Work