Affective Computing in Intelligent Vehicles

Learn how a car can process emotional data while you are driving and work to alleviate those stressors.

Hi, my name is Daria Loi, principal engineer at Intel Labs. Today, I will discuss how affective computing helps us develop intelligent vehicles. Stay tuned.

Affective computing is a capability by which devices understand human emotions and use that understanding to take actions. So what can be done with emotional data when you're driving? Let's look at an example. Bethany recently moved to the US. English is not her first language. Bethany is driving Melissa to her English lessons before heading to work, where she will present to her CEO.

Let's have a look at what technology is in Beth's car. Like all autonomous vehicles, her car is equipped with an electronic control unit in charge of its autonomous capability. In addition, the unit includes a system that monitors the driver and what they can provide passengers. Both systems share the same capabilities and leverage multiple sensors located inside the vehicle.

Let's now see what the occupant profiling system does in response to Melissa's emotional state. The occupant profiling system extracts features and determines that the child is agitated. After that, it pings Melissa's wearable device to verify whether she's agitated due to fatigue. Since fatigue or lack of sleep are not the cause for agitation, the system then looks into the child's calendar and discovers that she has an English class that day.

The affective logs also show a pattern. She's always agitated before [the] English class. The system records a new item in the affective log. On top of Melissa's agitation, Beth is going to be 20 minutes late to her presentation due to heavy traffic. Let's see how bad she's feeling. The driver monitoring system extracts a number of features and establishes that Beth is not only agitated, she's heavily distracted.

A smart vehicle at this point would take control, but this is an intelligent vehicle and it knows that it is unlikely Beth would let go of the steering wheel while distressed. She wants to feel in control and drive fast, possibly above the speed limits. So the system, instead, addresses her emotional state first. It notifies Beth's delay to the organizers, requests to adjust the agenda, and cues up the presentation.

Beth's heartbeat goes down. The system detects the she's no longer distracted. Already at this point, the vehicle offers to drive. Beth agrees, and starts rehearsing her presentation. Affective computing is not only about helping you avoid mistakes. It is also about enriching your life and focus on the things you care about.

Now that Beth's emotional state [has] improved, the system informs her about Melissa's state, providing a chance to tend her issues. Beth focuses on her daughter, and the system learns and logs something new about Beth's value system. Affective computing is the human side of artificial intelligence. A clear understanding of people's attitudes and thresholds related to AI will help us design meaningful systems that we will want in our daily lives.

Thanks for watching. You can access further resources on affective computing in the links below. Don't forget to like this video, and subscribe to the Intel® Software YouTube* [sic] channel, and check us out on Facebook*.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.