Case Study · Project 01

Takeover UX

Safety-Critical Interaction Design for L2/L3 Vehicles.

Automotive UXADASHuman-Centered InteractionMultimodal Interaction
Key Activities
Research · Synthesis · Interaction Design · Prototyping
Timeline
2 Weeks
Tools
Figma · Miro · Google Survey · Krea · Sketchbook · Alias
Focus Areas ·Safety-critical UX|Takeover readiness|Trust & urgency management
Takeover UX — cluster concepts: welcome screen, destination search, and active driving state with multimodal alert rail.
Live Instrument Cluster · Motion Study

The handover, in motion.

A continuous transition from automated driving to driver control.

Takeover UX · Cluster Motion
· 01 The Challenge

When the car
needs you back.

When an automated vehicle reaches its limits, a disengaged driver may have only seconds to respond. Poorly designed alerts can cause delayed reaction time, panic or confusion, and loss of trust in automation.

HMW

How might we support drivers in understanding when, why, and how to take control under time-critical conditions — without overwhelming them?

Goals
  • 01Enable fast and confident takeover
  • 02Reduce cognitive overload during stress
  • 03Communicate urgency clearly and progressively
Constraints
  • 01Time-critical scenarios (seconds matter)
  • 02High-risk environment
· 02 User & Scenario

A driver, a moment.

Target User

Driver using a partially or conditionally automated vehicle (L2/L3).

Context
  • ·Driver may be relaxed, distracted, or monitoring passively
  • ·Sudden environmental uncertainty (construction, weather, lane loss)
Core Scenario

The vehicle detects a situation it cannot handle and requests the driver to take over control safely and in time.

· 03 Design Process

From signal to system.

Step 01
Survey & Interview
User Survey
Step 02
Key Insights
Synthesis
Step 03
Persona
Driver archetype
Step 04
Story & Scenario
User Story
Step 05
Hierarchical Task Flow
Task mapping
Step 06
Wireframing & Hi-fi
Design execution
· 04 Research Methodology

Scenario-based user survey.

Research Scope

A nine-participant scenario survey informed this early design direction. The interface remains a hypothesis that needs simulator testing.

Because real takeover events are rare, nine participants responded to a scenario-based survey about expectations, concerns, alerts, and information needs. This was exploratory research—not usability testing or a controlled evaluation.

Participants had varied driving habits; 56% drove daily or almost daily, and 78% had limited or no automated-vehicle experience. Their responses exposed expectations and trust gaps among unfamiliar users.

Cohort
9
Participants
Cohort
78%
With limited or no automated-vehicle experience
Cohort
56%
Drive daily or almost daily
Key Insights
Insight 01

Early & Multimodal Alerts Are Essential

Visual + Audio + Haptic. Reduces missed alerts and supports divided attention.

Insight 02

Transparency Builds Trust

Users want to know why takeover is needed. Context reduces hesitation and doubt.

Insight 03

Clear Action Guidance Matters

Unclear actions increase stress. Users prefer narrowed, directive guidance.

Insight 04

Preference for Gradual & Slow Alerts

Strong preference for early warnings. Time to mentally prepare matters.

· 05 Persona

Meet Alex.

“I'm okay taking control… but don't surprise me or leave me guessing.”

A driver who accepts automation but becomes anxious when a sudden takeover request lacks context, time, or clear action.

Alex
Age 25
Driving Frequency
Drives daily or almost daily
Autonomous Experience
Limited to L2/L3 ADAS, no control
Goals
  • +Retain control quickly and safely
  • +Avoid panic during takeover
  • +Understand why (context) is required
  • +Feel confident once control is transferred
Needs
  • ·Early or gradual takeover alerts
  • ·Clear explanation of the takeover request
  • ·Simple, increased action guidance
  • ·Confirmation of manual control
Pain Points
  • −Anxiety caused by uncertainty
  • −Increased cognitive load under time pressure
  • −Hesitation due to multiple possible actions
  • −Loss of focus when system behavior fails
Frustrations
  • −Sudden, alarming takeover alerts
  • −No context for why takeover is happening
  • −Vague instructions like "Brake or steer as soon as you see alert"
· 06 Design Principles

Six rules for a calm handover.

Principle 01

Progressive urgency over sudden alerts

Supports emotional readiness under stress.

Principle 02

Clarity over completeness

Reduce cognitive load during takeover.

Principle 03

One dominant action at a time

Prevent hesitation and decision paralysis.

Principle 04

Multi-modal communication for redundancy

Ensure alerts are noticed across attention states.

Principle 05

Close the loop after control transfer

Reinforce trust and perceived safety.

Principle 06

Fail-safe design

Design for edge cases to ensure additional safety & user trust.

· 07 Solution

A multimodal, phased takeover.

A phased sequence prepares the driver, escalates urgency, and confirms control transfer.

Core Interaction Flow — phased takeover flow designed to reduce surprise, support readiness, and ensure a confident transition of control.
Step 01

Low vehicle engagement

System detects driver disengagement.

Step 02

Initial warning

Visual + Audio cue. Gentle reminder.

Step 03

Takeover request

Escalating urgency.

Step 04

Driver response

Driver takes control.

Step 05

Confirmation

System confirms successful takeover.

Step 06

Post-takeover support

Reassurance and feedback.

· 08 Low-Fidelity Interaction Explorations

Exploring six concepts, refining one direction.

Monitoring
Concept 01
Concept 01

Vertical Alert

Concept 02 — Alert on Vehicle Capsule
Concept 02
Concept 02

Alert on Vehicle Capsule

Concept 03 — Horizontal Alert Banner
Concept 03
Concept 03

Horizontal Alert Banner

Concept 04 — Icon-Rich Alert Panel
Concept 04
Concept 04

Icon-Rich Alert Panel

Concept 05 — Expanded Overlay Alert
Concept 05
Concept 05

Expanded Overlay Alert

Concept 06 — Compact Status Pill
Concept 06
✓ Chosen Direction

Compact Status Pill

Decision Making

I chose Concept 6 because its compact status pill preserves vehicle context while keeping the driver's attention on the road.

· 09 High-Fidelity Final UI

Take over immediately.

Low Visibility AheadPrepare To Take ControlTake Over ImmediatelyYou're In Control
State 01 · Monitoring

Low Visibility Ahead

System Monitoring Conditions

  • Calm blue ambient lighting signals passive system awareness without alarm.
  • Full driving information remains visible — speed, rpm, fuel, navigation.
01 / 04
Detail 01

Primary alert text above vehicle info

Preserving hierarchy so the directive reads first, vehicle data second.

Detail 02

Peripheral light panels

Light panels activate on the left and right, engaging peripheral vision.

Detail 03

Reduced info at final alert

In the final alert only speed is shown to the user to reduce information overload.

Detail 04

Progressive attention scaling

Color transitions communicate system state without abrupt visual disruption.

· 10 Outcome, Limitations & Next Steps
"The exploratory findings informed a hypothesis that progressive, multimodal communication may help reduce surprise and improve driver understanding during a takeover request. Controlled simulator evaluation is needed to test this hypothesis."

What was designed

A progressive four-state takeover interface — monitoring, prepare, takeover, in-control — coordinating cluster visuals, peripheral light, sound, and haptic cues into a single escalation language.

What was learned

In a nine-participant scenario survey, respondents described wanting advance warning, a clear reason for the request, and confirmation that control had transferred. Most had limited or no direct experience with automated vehicles.

What remains unvalidated

The interface has not undergone controlled simulator or usability evaluation. Comprehension, reaction time, workload, trust, and takeover quality are untested and remain open questions.

Future Evaluation

A next phase would test the concept in a driving simulator against a baseline alert, measuring takeover comprehension, reaction time, self-reported workload, trust in automation, and the quality of the resulting takeover manoeuvre.

© Prize T Mathew
· Academic Recognition

ACM AutomotiveUI 2026 · Student Research Track

An expanded version was accepted to the ACM AutomotiveUI 2026 Student Research Track for its design-led view of progressive urgency and multimodal takeover communication.

Acceptance reflects the contribution of the design perspective; it is not evidence that the interface is safer or more effective than an alternative.

· 11 Prototype Interaction Video

The prototype, in motion.

© 2026 Prize T Mathew
UC DAAP · MDes · Automotive UX/HMI