The Loss of Touch in Automotive HMI
UX Design
B.A Thesis | 2026
Overview
Touchscreens took over the car. Buttons disappeared. Nobody really asked whether that was a good idea.
The automotive industry's shift toward full-glass interiors was driven by aesthetics and cost reduction. The result is a generation of vehicles where adjusting the temperature, changing a track, or activating a wiper requires a driver to look away from the road, navigate a menu, and tap a flat surface with no physical feedback confirming the action landed.
This thesis argues that the problem was never the screen. It was the loss of touch.
The brief
Three speculative artefacts exploring what automotive HMI could look like if tactile usability were treated as a non-negotiable design requirement — not a nostalgic preference, but a safety and usability case.
The goal was not to remove screens from cars. It was to start a conversation about what gets lost when physical touch disappears entirely, and to propose design directions that bring it back — on glass and off it.
The research
Literature Research began in existing literature — papers on cognitive load theory, driver distraction, and the human factors behind road accidents. HUD projection limitations were examined alongside EU market data on car usage patterns and commuting behaviour. The goal was to establish what was already known about attention, interruption, and the cost of looking away from the road before touching any vehicle or talking to any driver.
Two frameworks were used to evaluate HMI design throughout the research:
De Waard's 1994 Cognitive Load Model, providing a structured way to assess the mental demand placed on a driver by any given interface interaction.
Don Norman's Affordance Framework: examining whether interface elements communicate how they should be used, and whether the feedback they provide confirms that an action has been completed.
These two lenses became the basis for evaluating every vehicle and every design decision.
The market research focused on the European automotive landscape: best-selling cars by segment, usage patterns, and commuting behaviour across the EU. This established the context for who is actually driving what, and on what kinds of journeys.
Market Research: Seven vehicles were selected across the budget-to-premium spectrum, representing the best-selling cars in Europe in 2025: Dacia Sandero · Renault Clio · Toyota Yaris · VW T-Roc · Audi Q5 · Mercedes GLC · BMW X3
POV driving footage was reviewed for each vehicle alongside user manuals. A modality map database was built in Notion, cataloguing every control surface and interaction type, then evaluated against the judgement framework.
The finding was clear: budget vehicles retained significantly more tactile controls. The shift from mid-range to premium was where screens began to dominate, and where tactile interaction was most aggressively replaced by glass. A telling postscript: during the writing of this thesis, Renault launched a facelifted Clio with a significantly more screen-dominant interface — moving it closer to the Q5 in interaction model. The trend is accelerating even in the budget segment.
Seven user interviews were conducted across five countries — Germany, Latvia, Lithuania, Poland, and Italy — with participants ranging from 3 to 24 years of driving experience. The interview group covered high-interruption short-distance commuting through to long drives exceeding 80km, and vehicles ranging from a Peugeot 107 to a Mercedes GLC to a Tesla Model 3.
The pattern that emerged was consistent: drivers of premium and screen-dominated vehicles reported the most frustration and the most frequent instances of inconvenience. The more the interface relied on touchscreen interaction, the more drivers felt the burden of it.
Convergence & Findings
Across both the vehicle analysis and the user interviews, one finding converged from both directions independently:
Steering wheel controls were the most used, most appreciated, and least contested interface element in the car. Drivers across all segments and experience levels would not remove them with one consistent modification request: make the buttons physical and tactile, not capacitive touch.
Capacitive touch on steering wheels (buttons that activate on skin contact without a click) was a consistent frustration. Drivers wanted to feel the button move. They wanted confirmation that didn't require looking.
Artefact Requirements
Research findings from the vehicle analysis and user interview transcripts were synthesised into a structured set of design requirements using Claude as a processing tool — the full dataset was uploaded and a requirements list was generated from the qualitative data, then reviewed, validated, and refined against the research frameworks.
Key requirements that shaped the artefacts:
-
Retain steering wheel controls; make them physical and tactile, not capacitive
-
Move high-demand functions to single-step physical buttons
-
Keep existing physical controls physical (do not replace them with screens)
-
Make high-frequency functions reachable in one step without menu navigation
-
Provide feedback that persists and can be confirmed without looking
-
Maintain a coherent, stable layout that doesn't change between interactions
Precedents
Tactus Technology developed a microfluidic touchscreen overlay for smartphones that raised physical bubble buttons on demand from beneath the glass surface. The technology worked. It failed commercially due to manufacturing costs and the industry's momentum toward flat glass. It remains the closest real-world precedent for on-demand tactile touchscreen interaction — and the starting point for the third artefact in this project.
Artefact 01. Steering Wheel
Physical, tactile button placement mapped to task class and interaction frequency. High-demand functions — those a driver uses most and needs most reliably — brought to single-step physical controls within thumb reach. Capacitive touch replaced with tactile buttons that click, confirm, and can be operated without eye contact.
Artefact 02. Dashboard
A rearranged centre stack and instrument cluster prioritising physical controls for high-frequency functions and reserving the screen for low-demand, low-frequency interactions. Layout coherence maintained so muscle memory can develop — the same control is always in the same place.
Artefact 03. Microfluidics On-Demand Touchscreen architecture
A speculative interface layer built on the Tactus Technology precedent — a microfluidic system beneath the touchscreen glass that raises physical button surfaces on demand, then retracts them when not needed. The screen remains a screen. But when a driver needs tactile confirmation, the surface gives it to them.
This artefact is speculative. The technology has proven precedent. The barrier is manufacturing cost and industry will, not physics.
Prototyping
Both artefacts were tested through physical prototyping to validate the tactile logic — not the engineering, but the interaction feel.
Lo-fi prototype: cardboard, foam tape, cling film, electrical tubing. Built to test button placement, reach, and the basic sensation of a raised surface under glass.
Hi-fi prototype: shrink paper, UV resin, silicone. Built to test the material behaviour of a surface that could plausibly raise and retract, and to evaluate what on-demand tactility actually feels like in the hand.
The prototypes work as interaction models. They confirm that the tactile logic is sound, that a raised surface under glass is findable without looking, and that the button placement tested in the lo-fi model translates to the hi-fi version. Engineering conclusions are beyond the scope of a design thesis, and that boundary is acknowledged honestly.
Limitations and Reflection
This is a speculative design project. The artefacts are not engineered for production, and no claim is made that they are. The microfluidic touchscreen in particular sits at the boundary of what a design student can physically prototype, the concept has real technological precedent, but the manufacturing pathway is an engineering problem, not a design one.
Time constraints limited the depth of prototyping and the number of user interviews. Seven interviews across five countries is a meaningful sample for a thesis project; it is not a statistically conclusive dataset. The goal of this project was never mass production. It was conversation about what the industry has traded away in pursuit of the clean glass interior, and whether drivers actually asked for that trade.
The automotive industry moved fast toward screens and called it progress. This thesis tried to slow that assumption down long enough to ask who it was progress for. The research kept returning to the same answer: not the driver. The driver lost something when the buttons disappeared — not just convenience, but the kind of quiet confidence that comes from knowing exactly where your hand is going and feeling it arrive. That confirmation matters at 120km/h on a motorway. It matters more than a clean dashboard.
The three artefacts don't solve the problem. They propose a direction and make the conversation harder to avoid. That felt like enough for a thesis. Whether it becomes enough for the industry is a different question.
[other roles]
