Immersive Emotion Induction: Beyond Static Images to Interactive Driving Simulators

TL;DR

Researchers from Hyundai Motor Group, Korea University, and Fraunhofer IAO have developed a novel paradigm for inducing specific human emotions—such as anger, boredom, and trust—using immersive driving scenarios. By replacing static images with interactive, high-stakes virtual environments, they have established a more effective framework for evaluating Human-Computer Interaction (HCI).

The Immersion Gap: Why Static Stimuli Fail

In the world of psychology and HCI, the International Affective Picture System (IAPS) has long been the gold standard. However, looking at a picture of a car crash is fundamentally different from a participant experiencing a near-miss in a simulator.

The core problem identified by the authors is the Lack of Immersion. Traditional methods provide a passive experience where the "subject" is decoupled from the stimulus. This paper argues that emotion is not just a reaction to an image, but a reaction to an interaction. In a driving context, emotions are high, stakes feel real, and the sensory input is continuous—making it the perfect laboratory for affective computing.

Methodology: Specificity through Scenarios

The researchers didn't just put people in a car; they designed "Scenario-Specific Induction." Instead of looking for broad happiness or sadness, they targeted emotions relevant to modern technology adoption:

1. Anger/Frustration: Induced by a scenario where a truck consistently blocks a highway overtaking maneuver.
2. Boredom: Triggered through long, monotonous driving without the ability to overtake.
3. Satisfaction/Dissatisfaction: Manipulated through throttle response (acceleration speed) when following other vehicles.
4. Trust in Technology: Measured specifically during an Automatic Emergency Braking (AEB) event.

System Architecture

The hardware setup provides a 180-degree frontal view using three planar screens to simulate front and side mirrors, creating a high level of situational awareness.

Multi-Modal Verification

How do we know the participants actually felt these emotions? The study employed a triangulation of data:

• Physiological Sensors: Heart Rate (HR) and Galvanic Skin Response (GSR).
• Behavioral Data: Facial expression tracking using the Facet algorithm.
• Subjective Feedback: Likert scales and visual analogue questionnaires.

This multi-dimensional approach allows researchers to observe the mismatch or alignment between what a user says they feel and what their body actually does—a critical insight for developing safe and intuitive vehicle interfaces.

Results & Critical Insight

The study successfully demonstrated that interactive scenarios can consistently induce targeted emotions. The key takeaway for the industry is that Context is King.

While a static image of a snake might induce fear universally, the emotion of "distrust" in a self-driving car can only be measured when the user is placed in a scenario where the car must act autonomously under pressure. The results showed significant physiological shifts during the AEB and "hindered overtaking" segments, proving the paradigm's efficacy.

Limitations & Future Work

As a SIGGRAPH Asia poster, the depth of the longitudinal data is limited. Future research would benefit from:

• Cross-Platform Validation: Testing if these scenarios are as effective on HMDs (VR headsets) compared to large-scale simulators.
• Individual Heterogeneity: Analyzing why some "drivers" were more susceptible to anger than others under the same scenario conditions.

Conclusion

This work represents a vital shift towards Active Affective Computing. By moving the laboratory into a virtual, interactive world, we can better understand how human emotions will dictate the success of future technologies, from AR interfaces to fully autonomous vehicles.