Usability Testing of Tesla Model 3 Digital cockpit
My Role : Problem Identification | Usability Testing
Duration : 8 weeks
Team : Yavnika Miglani | Prerana Jayakumar | Lauren Hoese
Project Overview
This academic project focuses on identifying and addressing usability challenges in the Tesla Model 3’s digital cockpit. Through a detailed literature review, we analyzed existing issues related to the accessibility, clarity, and efficiency of the in-car display system. Based on the insights gathered, we developed and tested a redesigned prototype aimed at improving the overall user experience, enhancing safety, and supporting greater personalization within the digital cockpit environment. The study highlights the importance of user-centered design in automotive interfaces and explores how thoughtful redesigns can mitigate safety risks while enhancing driver satisfaction.
PROBLEM IDENTIFICATION
Cluttered and Confusing Icon Layout
Icons were too small, visually similar, and densely packed without clear labeling, making it difficult for users to quickly locate and interact with essential functions.
Inconvenient Access to Frequently Used Mobile Applications
Core applications required deep navigation into multiple sub-menus, increasing cognitive load and distracting users from the primary driving task.
Ambiguous Driving Mode Status
The active driving mode (e.g., Autopilot, Manual) was not clearly communicated, creating uncertainty and potential safety risks during operation.
SOLUTION
Solution 1
Increased the icon size for better touch accuracy, added clear text labels for easier recognition, and removed or repositioned redundant icons to declutter the main interface.
Solution 2
Introduced a frequently used features drawer for quick access from the main screen.
Solution 4
Applied color differentiation to driving modes : blue for manual driving and magenta for autopilot to make the current mode instantly recognizable at a glance.
Solution 3
Added personalized profiles with a valet mode to customize the interface based on different driver preferences and usage needs.
USABILITY TESTING
Equipment
Hardware
Software
SMI Binocular Eyetracking Glasses , SMI Dongles , Logitech steering wheel console Ipad,
Figma, Excel, Notion, Google Docs
Participant Recruitment
6 participants
Diverse in age, gender and driving experience
Mixed familiarity with ADAS technology
No experience with Tesla cars
Testing Scenerio
Simulation Setup
Testing Method
Goals
Post Evaluation
Logitech steering wheel + Slowroads driving simulator
Tablet prototype placed beside the user as the in-car display
Gaze-tracking glasses worn to monitor attention and interaction patterns
Wizard of Oz technique
A team member secretly controlled the system’s responses like autopilot feature.
A second facilitator guided the user through tasks (navigation, feature testing)
Shown actual Tesla Model 3 interface
A/B comparison with our prototype
Collected user feedback on clarity, preference, and overall experience
Test usability, personalization, and natural interaction of the display
Measure speed, accuracy, and ease of use in a realistic context
Task 1 (Switching between Autopilot and Manual Mode) :
Task 2 (Accessing Notifications and Playing Media):
Task 3 (Launching an App Using the Frequent App Drawer):
Task 4 (Adjusting In-Car Settings):
Benchmark set at 2 seconds, as quick decisions are crucial to avoid distraction. Research shows that tasks exceeding this duration can heighten crash
Benchmark set at 5 seconds, balancing the need for interaction with minimizing risk. This duration is based on guidelines for secondary tasks in vehicles.
Benchmark set at 5 seconds due to the complexity, but anything longer could dangerously divert attention.
Benchmark set at 3 seconds, ensuring the task is quick enough to prevent significant distraction.
Tasks and Benchmarking
To determine whether the time taken by participants for each task is within acceptable limits for a safe driving user experience, we needed a benchmark for task completion times. According to research, tasks that require more than 2 seconds of visual attention away from the road significantly increase the risk of accidents .
Data from surveys were analyzed using heat maps produced from eye gaze tracking videos. The objective was to identify significant relationships between personalization features and user experience metrics, preferences, and user biases observed during testing.
Quantitative Data Analysis
Task 1 (Navigating, using Autopilot, and switching back to Manual mode):
Observation: Users easily engaged the Autopilot but struggled to switch back to Manual mode, repeatedly pressing the Autopilot button instead of using the provided manual switch.
Heatmap Analysis: The heatmap showed concentrated gaze on the Autopilot button, indicating confusion or lack of awareness of the manual mode switch.
Task 2 (Accessing and responding to notifications, playing media):
Observation: Participants found the notification drawer easy to access and appreciated the quick action buttons.
Heatmap Analysis: Smooth gaze patterns indicated a straightforward interaction with this feature.
Visualized Eye Tracking data for Task 2
Visualized Eye Tracking data for Task 1
Task 3 (Launching an app using the Frequent app drawer):
Observation: This task was performed with ease, with participants quickly accessing the app drawer without distraction.
Heatmap Analysis: Minimal gaze deviation showed participants were not confused and could focus on the task.
Task 4 (Adjusting in-car settings):
Observation: This task was the most challenging. Participants had difficulty with scroll functions and were hesitant to click on the wrong buttons.
Heatmap Analysis: Erratic gaze patterns indicated confusion and uncertainty.
Time Taken for Each Task
Participant | Task 1 (sec) | Task 2 (sec) | Task 3 (sec) | Task 4 (sec) |
|---|---|---|---|---|
P1 | 3.8 | 13.2 | 3.9 | 15.3 |
P2 | 4.1 | 12.8 | 4.2 | 14.7 |
P3 | 3.9 | 13.4 | 4.1 | 15.1 |
P4 | 4.0 | 13.1 | 4.0 | 14.9 |
P5 | 4.2 | 13.0 | 3.8 | 15.4 |
P6 | 3.9 | 13.3 | 4.0 | 15.0 |
Average | 4.0 | 13.1 | 4.0 | 15.1 |
Comparison with Benchmark:
Task 1: Average time (4.0 seconds) exceeded the benchmark by 2 seconds, indicating that switching between Autopilot and Manual mode might be too complex.
Task 2: Average time (13.1 seconds) far exceeded the benchmark, suggesting the need for quicker access or less visual distraction.
Task 3: Average time (4.0 seconds) exceeded the benchmark by 1 second, indicating that the app drawer could be more intuitive.
Task 4: Average time (15.1 seconds) significantly exceeded the benchmark, showing that adjusting in-car settings requires too much attention and is a potential safety risk.
Visualized Eye Tracking data for Task 3 & 4
Eye-Tracking and Gaze Data Analysis
Benchmarking Tasks
Task 1 (Switching between Autopilot and Manual Mode): Benchmark set at 2 seconds, as quick decisions are crucial to avoid distraction. Research shows that tasks exceeding this duration can heighten crash risk.
Task 2 (Accessing Notifications and Playing Media): Benchmark set at 5 seconds, balancing the need for interaction with minimizing risk. This duration is based on guidelines for secondary tasks in vehicles.
Task 3 (Launching an App Using the Frequent App Drawer): Benchmark set at 3 seconds, ensuring the task is quick enough to prevent significant distraction.
Task 4 (Adjusting In-Car Settings): Benchmark set at 5 seconds due to the complexity, but anything longer could dangerously divert attention.
Qualitative Data Analysis
Interview transcripts, videos, feedbacks, and eye gaze tracker videos were all analyzed in order to identify and understand recurring themes and patterns in user feedback. The outcomes of expert interviews were cross-checked with user data in order to validate conclusions.
Key Feedback Themes
Usability Concerns: Participants frequently mentioned that the UI felt overwhelming initially, which could lead to distractions while driving. They recommended a more simplified version of the interface to reduce cognitive load and improve usability. Additionally, the absence of clear labeling on some features was noted as a challenge that could contribute to a steep learning curve.
Visual Design Preferences: The default dark mode raised issues related to readability and contrast. Participants suggested that more commonly recognized icons and labels could improve ease of use. There was also a desire for interaction patterns that minimized the need for users to divert attention from driving, such as a more consistent and intuitive scrolling or clicking mechanism.
Customization and Accessibility: Although participants appreciated the larger icon sizes, they expressed a desire for more customization options, particularly in terms of icon placement and feature accessibility. Enhancing the prominence and customizability of the frequent app drawer was also suggested to improve both safety and user experience.
Behavioral Observations
During interaction with the UI, participants often slowed down or stopped, which could pose safety risks, especially in high-traffic or highway scenarios. Some users described the interface as overly detailed, likening it to a computer desktop rather than a typical car UI. This level of detail was perceived as overwhelming, potentially leading to hesitation or confusion. Additionally, difficulties in quickly locating key features caused frustration and delays, highlighting the need for more intuitive and accessible design elements.
Discussion & Conclusion
The testing revealed that while the Tesla Model 3's redesigned UI offers extensive features, it may be overly complex for safe and intuitive use during driving. The task completion times exceeded the recommended benchmarks, indicating that the interface requires significant refinement to reduce cognitive load and enhance safety. Simplifying the UI, improving visual contrast, and integrating more non-visual controls (like haptic feedback) are recommended steps moving forward.