Nova - Redefining The Relationship Between Driver & Vehicle.
Nova AI is an intelligent IVIS co-pilot designed to bridge the gap between high-tech automation and road safety. By transforming the In-Vehicle Infotainment System (IVIS) into a proactive partner, Nova utilizes predictive assistance and context-aware interactions to slash cognitive load. From preempting mechanical failures to adaptive rerouting, Nova ensures that while the car gets smarter, the driver stays focused on what matters most: the road.
Problem Statement
Driver interaction with IVIS systems increases distraction and cognitive workload. Complex navigation tasks demand sustained attention, while delayed responses to accidents or mechanical failures can result in life-threatening consequences. AI-based personal assistants offer potential solutions by enabling predictive assistance, context-aware interaction, automated emergency detection, and adaptive rerouting.
Drivers are often bombarded with non-essential data, leading to "glance fatigue," which increases the risk of lane deviation and delayed braking.
When a car suddenly suggests a new route or flags a mechanical issue without context, leads to system abandonment or distrust.
"Engine has some issue" is indicated with just an icon. For normal users, this information is stressful and difficult to interpret. Exactly what is the issue?
UX Road Map
Our roadmap was divided into a three-phase User-Centered Design (UCD) process:
Emphatize Phase
In this phase, insights were synthesized from user surveys, observational task analysis, and 15 research articles focused on AI trust, cognitive workload, and multimodal ergonomics.
Latency & Reliability:
Multimodal sensing requires sub-5-second processing to be effective in emergency scenarios.
Human-in-the-loop:
Automated actions (like emergency calling) require a confirmation/cancel loop to ensure driver agency.
Information Density:
Audio-first alerts are critical for significant traffic changes to prevent visual "glance fatigue".
The "Why" Matters:
Drivers explicitly need trusted AI guidance that provides simple explanations alongside route suggestions.
Multimodal Clarity:
There is a strong demand for clear, combined audio and visual reroute alerts.
Control and Goals:
Drivers are primarily motivated by saving time and avoiding stress. They want options to choose between the fastest, safest, or fairest routes to reach their destinations reliably.
Key Pain Points:
Drivers reported immense frustration with delayed traffic updates, inaccurate ETAs, intrusive alerts that overwhelm them while driving, and abruptly suggesting unfamiliar reroutes without offering any transparent reasoning.
While users prefer typing on their phones before a trip, they switch entirely to voice commands while in motion.
Observation: The most dangerous cognitive spike occurs when drivers attempt to evaluate and change routes mid-trip, highlighting a critical need for low-effort, transparent rerouting assistance.
Communication Friction
Current voice assistants frequently fail to recognize specific contact names (especially those saved with emojis), forcing drivers to manually scroll through digital phonebooks.
Observation: This creates a dangerous loop: vehicle vibrations cause screen misclicks,and drastically increasing visual distraction from the road.
Touchscreen-heavy interfaces for basic functions like climate and media control require precise tapping, dangerously pulling the driver's eyes away from the road.
Observation: Drivers frequently made adjustments and strongly prefer the tactile feedback of mechanical steering wheel buttons or highly reliable voice inputs over navigating deep touchscreen menus.
Affinity Diagram
User Persona
Ideation & Prototyping Phase
Low fidelity Wireframe
Persistent Control Panel:
Based on task analysis and user feedback during low-fidelity wireframe iteration, frequent actions like climate control, music, and volume were anchored as a persistent layer on the display, rather than being buried in sub-menus.
Proximity-Based Layout:
Highly interactive widgets (like navigation) strictly on the left side of the screen, ensuring the closest physical reach for the driver during high-stress or emergency situations.
Dedicated AI Information Zone:
A specific, consistent area of the screen solely for AI alerts and explanations appears from the right side of the IVIS, training the driver's eye exactly where to look when the system initiates a prompt.
High-fidelity Prototype
Dark-Mode Default:
We utilized a dark, high-contrast color palette as the default interface to minimize screen glare in the cabin and reduce driver eye strain across varying lighting conditions.
Color-Coded Urgency:
We implemented a strict traffic-light color system for AI alerts: Green for standard reroutes, Yellow for predictive mechanical warnings (e.g., tire pressure drops), and Red for critical emergencies (e.g., collision detected).
Multimodal Feedback:
Visual elements were tightly synchronized with voice prompts. The screen visually highlighted exactly what the AI was stating aloud, ensuring the driver received redundant, clear signals without needing to stare at the display.
User Testing Phase
Users were tasked with navigating three critical scenarios:
1] A minor accident detection
2] Adaptive rerouting due to traffic
3] A mechanical diagnosis
Method
The "Wizard Of Oz" method was used, where the instructor imitated an AI.
A "think-aloud" procedure followed by a subjective usability questionnaire and the System Usability Scale (SUS).
Multimodal Redundancy:
The NOVA AI system effectively addresses the paradox of modern infotainment by shifting from a passive interface to a proactive co-pilot. By prioritizing transparency and multimodal collaboration, cognitive load was successfully reduced while maintaining the safety and agency of the driver.
Advanced Personalization: