Research Highlights
To develop a workflow capable of lab-based controlled data collection
September 04, 2025

The Science Objective
To develop a workflow capable of collecting first-person-perspective data in a controlled, laboratory-based environment
Approach
- We adopted VR as the main immersive and controlled laboratory method. Using the 3D Gaussian Splatting technique, we reconstructed photo-realistic scenes from the real world to serve as the foundational materials for subsequent VR simulations.
- Semantic segmentation, text-image AI, and VR component development were also utilized in the customized editing and data recording processes.

Impact
- The study improved the traditional VR modeling and editing workflow, offering significant advancements in both visual realism and cost-efficiency for immersive environmental perception research and human-scale data collection.
- Moreover, it can aid policymakers and urban designers in exploring human-perspective evidence to better improve the urban built environment, encourage active travel, and promote transport decarbonization.
Summary
The project introduces a new workflow for active mobility data collection using VR and 3D Gaussian Splatting to create photo-realistic urban scenes. It integrates semantic segmentation, AI editing, and VR development to improve visual realism and cost-efficiency. The approach enables first-person perspective data in controlled settings, supporting research, policymakers, and urban designers in promoting active travel and transport decarbonization.
Team Members

Yangbo Bi (Postdoc)
University of Birmingham

Phil Jones
University of Birmingham
Publications
TBS
Sign up for our mailing list
You’ll receive the latest updates on CLEETS research and programming.