Hyundai IONIQ 6 — designing for trust at L2+

How might we increase driver confidence in Hyundai's HDA 2.0 system under semi-Level 2 and Level 2+ autonomy — where control quietly slides between human and machine, and trust is the variable that decides everything?

Nine drivers in real and simulated conditions. ECG (heart-rate variability), Tobii eye tracking, and self-report Likert surveys. A custom Unreal Engine simulator with surround sound and highway rendering. A/B UI testing, paired-sample t-test, and a confidence metric synthesized from biometric and behavioral data.

Less information builds trust in partial autonomy; more information reassures in full autonomy. The dose changes with the autonomy level — and the existing HDA UI didn't titrate, so every handoff cost the driver something.

White UI as a minimal baseline. Green UI as an emotionally calming, trust-oriented direction. Blue UI as a highly technical, information-rich variant. Each was scored on the confidence metric and triangulated against driver feedback.

Green UI achieved a +10% improvement in confidence score over the baseline. The final solution layered dynamic icon feedback (paused vs active), adaptive info density tied to driving mode, high-contrast lane keep + distance follow alerts, and a steering-wheel-mounted screen as the North Star concept.

Simulator integration so users can learn the system's behaviors before trusting them. Complex cityscape testing to extend findings beyond highway. Stratified participant selection by Hyundai-vehicle familiarity to isolate prior bias from new design effect.