MobiUX
Visuals & Confidentiality
As this project is ongoing and governed by strict ethical guidelines and confidentiality agreements, images and detailed visuals of the simulator cannot be shared at this time.
Here are some older photos from the early stages of the simulator development:
Designing MobiUX: A Driving Simulator for Safer, Smarter Detection
Simulating real-world driving behavior to build safer AI systems
As Project Lead and UX Designer, I led the development of MobiUX, a modular driving simulator designed to train and validate driver impairment detection systems.
What began in Japan as a minimal training rig evolved under my leadership into a full-cabin simulator in Michigan. We transformed early insights into a powerful research tool capable of collecting nuanced human-machine interaction data across varied demographics and environments.
This simulator now serves as a cornerstone in Mitsubishi Electric’s in-cabin AI research ecosystem.
Role & Responsibilities:
Simulator UX • Experiment Design • System Integration • Cross-functional Leadership
Tools:
Figma, Unity, Python (data logging), Heart Rate Monitors, DMS Cameras
Duration:
2022 - Present
Team:
AI Research, Human Factors, HMI, Safety & Ethics, Engineering (Japan & U.S.)
01/ From Concept to Cabin
Japan (Prototype Phase):
-
Small-scale setup: screen, pedals, DMS camera, seat, heart rate monitor
-
Used for initial algorithm training on distraction and impairment
Michigan (Full Platform):
I led the design and build of a fully modular cabin with:
-
6 seats (4 removable for 2-seat config)
-
Steering wheel, pedals, gear shift, surround sound
-
120” screen, haptic feedback, and ambient realism
-
7-camera cabin array, DMS, and heart rate sensors
-
Functional infotainment touchscreen running FLEXConnect™
02 / Experiment Protocol
To validate impairment detection, we developed a rigorous and ethically managed test process:
-
Participants consumed alcohol in a controlled setting
-
Drove through custom scenarios (e.g., urban routes, child pickup, parallel parking)
-
We tracked:
-
Vitals via heart rate monitor
-
Gaze, eye closure, posture, and distraction via DMS
-
Driving behavior (acceleration, braking, steering)
-
-
Data captured in real time across varied impairment levels
-
UX designed to ensure smooth protocol flow, safety monitoring, and session repeatability
03 / System Components
Hardware + Sensors:
-
Full automotive cabin shell
-
6 modular seats, real steering controls, gear shift
-
120” projection screen with wraparound visibility
-
Surround sound + haptics for realism
-
7-camera in-cabin vision system
-
DMS + heart rate sensor
-
Touchscreen infotainment with FLEXConnect UI
UX Features:
-
Operator dashboard for live session monitoring
-
Scenario flow design: time-based tasks,
-
performance cues
-
Real-time alerts for vitals or behavior outliers
-
Full data tagging and logging integration
04 / Highlights & Contributions
-
Project Lead overseeing all phases from concept to deployment
-
Directed simulator hardware selection, team coordination, and cross-site integration
-
Created a modular, transportable simulator for use in AI, UX, and safety testing
-
Designed an experiment protocol that enabled real-world biometric and behavioral testing
-
Led interface design to support researchers, participants, and algorithm developers
-
Enabled higher accuracy in impairment detection across a broader participant demographic
05 / UX Design Approach & Learnings
-
Human-Centered Simulation: Designing around impaired behavior meant prioritizing realism, empathy, and observability
-
Multi-Stakeholder UX: Interfaces had to support not only drivers but also researchers, data analysts, and safety observers
-
Scalability & Modularity: System design emphasized ease of reconfiguration, transport, and future upgrades
-
Cross-Cultural Collaboration: Bridged work between Japanese research foundations and US-based hardware development
For further details regarding my work on this project, please contact me on sanajafri7@gmail.com