Agentic HMI for Autonomous Cars
Industry
Automotive
Expertise
UX - Unity Development
Tech Stack
Unity . Claude Code . Gemini . Blender
Tech Stack
Unity . Claude Code . Gemini . Blender
A UX design study of a HMI for a Level-4 (highly automated) vehicle that learns from repeated driving situations. Because the car drives itself, the interface’s job shifts from instructing a driver to earning a passenger’s trust through transparent, adaptive communication. The worked example is a narrow construction-zone passage where the car must pass a vehicle in the adjacent lane with little lateral clearance and the HMI behaves differently the first time it meets the situation, after it has handled it many times, and when something unexpected breaks the pattern. A functional prototype was built in Unity on the Unity Automotive HMI Template.
Level-4 automation (SAE J3016) means that, within a defined operational design domain (ODD), the vehicle performs the entire driving task with no expectation of human intervention. The person in the seat is a passenger, not a driver. This inverts the role of the cluster HMI. It no longer tells a driver what to do; it keeps a passenger oriented, calibrates their trust, and explains the car’s intent. The design tension is communication volume: over-explaining every routine manoeuvre erodes confidence, while under-explaining a genuinely tricky moment erodes trust. A fixed amount of information is wrong in both directions — so the HMI must adapt. The chosen situation is a congested, narrow, under-construction stretch limited to 40 km/h, with a larger vehicle in the adjacent lane. It is a lateral-clearance problem — “can I pass without rubbing?” — exactly the kind of edge a passenger watches nervously, which makes it a good probe for trust and transparency.
Next Work