Introduction

Driving simulators are a standard tool in human factors research, but most are tied to one laboratory. That limits sample size, the diversity of participants and how quickly scenarios can change. These limits matter more as mobility research moves beyond the driver and the vehicle to the many road users who now share streets and sidewalks.

HRI MobiSim addresses this with a single Unity project that builds to two targets: a WebGL version that runs in a web browser for large crowdsourced online studies, and a native VR version for in-lab studies with physical hardware and physiological sensors. Surveys, attention checks, condition assignment and continuous behavioral logging are built in, so a study can move from online screening to in-lab replication without rebuilding the scenario.

At a glance

• Built in Unity. Scenarios are set up in the Unity Inspector, mostly without writing code.
• The participant can be a car driver, e-scooter rider, wheelchair user, pedestrian or delivery robot.
• Online studies run in the browser, with a lightweight PHP server that checks participant IDs and collects data. It works with Prolific and MTurk.
• In-lab studies work with PC-based VR headsets, physical controls such as a scooter handlebar, and physiological sensors.
• Used as the main apparatus in three peer-reviewed studies with 458 participants in total.
• Available for non-commercial use.

Key features

Deployment options

Online, in the browser (WebGL)

Best for large and diverse samples.

• Participants open a link from Prolific or MTurk in a desktop browser and use the keyboard.
• A small PHP server validates each participant ID, blocks repeat sign-ups and stores the data.

In the lab, in virtual reality

Best for high-fidelity studies and physiological measures.

• Works with PC-based VR headsets (tested with the Varjo XR-3). Our lab setup adds a physical scooter handlebar on a Fanatec base and a Shimmer3 wristband for heart rate and electrodermal activity.
• Experimenters can trigger scripted events and code behavior live from a tablet Remote Control app.

Both versions are built from the same project, with the same scenarios, agents, Scene Graph and log format, so results can be compared directly across settings.

Published studies

HRI MobiSim has been the main experimental apparatus in these peer-reviewed studies:

What’s included

• Full Unity project source code
• Server-side PHP scripts for participant authorization and data collection
• Sample scenarios and ready-made actors
• Online documentation, API reference and a browser demo (no installation or account needed)

System requirements

Citation

If you use HRI MobiSim in your research, please cite:

Salubre, K., Lo, H.-C., Akash, K., & Misu, T. (2026). Mobility Interaction Sim: A scalable experimental research platform for mobility research. Proceedings of the Human Factors and Ergonomics Society Annual Meeting. https://doi.org/10.1177/10711813261485186

@article{salubre2026mobisim,
  title     = {Mobility Interaction Sim: A Scalable Experimental Research
               Platform for Mobility Research},
  author    = {Salubre, Kevin and Lo, Hiu-Chun and Akash, Kumar and Misu, Teruhisa},
  journal   = {Proceedings of the Human Factors and Ergonomics Society Annual Meeting},
  year      = {2026},
  publisher = {SAGE Publications},
  doi       = {10.1177/10711813261485186}
}

Download the software

The software is available for non-commercial usage. You must be affiliated with a university and use your university email address to make the request. Use this link to make the download request.

Want to try it first? The browser demo needs no installation or account.