SIM-26 Driving Simulator
Undergraduate Research Winter/Fall 2026 Multidisciplinary Grade: #
Info Detail
------- -------
TIME Jan 2026 – Present
LAB/ORG Multidisciplinary Design Program @ U-M;
University of Michigan Transportation Research Institute
ROLE Technical Lead & Software Engineer
STACK C++, Unreal Engine 5 (Blueprints & Networking)
Mission
The SIM-26 project is a large-scale multidisciplinary research initiative at the University of Michigan aimed at studying driver workload and distraction through various factors.
As the Technical Lead for the “Wizard of Oz” subteam, I architected a high-fidelity, multiplayer UE5 environment where subjects follow a vehicle they believe is fully autonomous, but is actually controlled by a hidden researcher. This architecture is essential for eliminating social attribution bias. By maintaining the illusion of an AI agent, we ensure participants react naturally to driving hazards rather than second-guessing a researcher’s intent. It also serves as a resource-efficient alternative to developing complex autonomous agents from scratch, providing a highly reactive and realistic simulation environment that serves the broader research team.
Implementation
Multiplayer Architecture & Telemetry
- Listen-Server Framework: Engineered a custom UI for host/join functionality and a dynamic vehicle spawning system in UE 5 using Blueprints and C++ to synchronize the “Wizard” and the participant in a shared simulation space.
- Data Collection: Integrated a visual occlusion system mapped to the steering wheel, allowing the team to quantify visual demand by tracking the frequency and duration of manual glimpses the subject used to navigate relative to the lead vehicle.
Experimental Design & Environmental Logic
- Traffic Setup: Scripted ambient background traffic to prevent participant “zoning out,” while limiting sessions to 15 minutes to avoid cognitive fatigue and focus loss.
- Environment Setup: Designed a rural road environment to minimize environmental load and confounding variables. The map uses gentle curves to keep focus on the lead car while utilizing specific textures and road furniture (trees, signs) to ensure accurate optical flow for speed perception.
Challenges
Network & Engine Onboarding
Building simulation on UE 5 required a mastery of its Blueprint scripting and multiplayer networking architecture. Beyond the engine, I navigated complex hardware-level hurdles to ensure a seamless “Wizard of Oz” connection. This included managing firewall configurations, optimizing local network protocols (TCP/UDP), and troubleshooting hardware-specific connectivity issues across Ethernet to maintain the low-latency environment for real-time interaction.
Technical Leadership & Synchronization
As a first-time Technical Lead, my primary challenge was shifting from execution to delegation. I focused on architecting a development roadmap that allowed for parallel workflows, ensuring each subteam member had clear, actionable tasks while maintaining strict project deadlines. Additionally, I acted as the primary liaison between our subteam and the broader SIM-26 project, following standardized documentation and communication protocols to ensure our “Wizard” module integrated perfectly with the wider research system.
Reflection
Moving from a contributor on teams like MRover to Technical Lead was a big jump. I had to shift from just writing code to managing the “big picture”, like integrating everyone’s work, hitting deadlines, and keeping documentation clean. It taught me that leading a large team is as much about synchronization as it is about technical execution.
I plan to carry these simulation and networking skills into my future robotics research. The server architecture I built for the “Wizard” setup is a versatile foundation for any project requiring real-time interaction between devices and humans, allowing for faster prototyping without the overhead of physical hardware.