GREENTRAVEL VR : Technical Report

Le résumé fourni par la source

This Technical Report documents the technical development of the GREENTRAVEL Virtual Reality (VR) system, created by the Digital Geography Lab at the University of Helsinki as part of the European Research Council-funded GREENTRAVEL project (ERC, 101044906). The report aims to provide a clear guideline for other research groups interested in applying VR and so help enhance the replicability and transferability of VR methods in urban planning research. The GREENTRAVEL VR system was designed to simulate realistic cycling experiences for users to test the psychological and physiological effects of different seasonality and street greenery scenarios in an urban environment. The system architecture integrates multiple hardware and software components for real-time interactive and immersive experiences. Navigation is achieved using a physical bicycle equipped with a Bluetooth Low Energy cadence sensor, which translates pedalling data into virtual movement. Visualisation and integrated eye tracking data capture are provided by the high-quality Varjo XR3 Head Mounted Display. The software application was developed in Unreal Engine, utilizing its visual scripting system, Blueprints, alongside C++ for complex functionalities. A custom-built WebSocket network module manages real-time data exchange, allowing synchronisation of events with physiological measurements, such as Electrodermal Activity. The design developed five distinct virtual environments: one neutral baseline and four street intervention scenarios (Built, Low level of greenery, High level of greenery and Winter) that share the same street geometry and built structure but vary in vegetation and seasonal conditions. A primary technical consideration was performance optimisation to mitigate cybersickness. This required continuous effort to ensure high and consistent Frame Rates, typically capped at 45fps for smoothness. To manage the high computational load associated with detailed 3D assets and hyper-realism, optimisation techniques such as Level of Detail, Distance Culling, and efficient Instancing were crucial for foliage, buildings, and street geometry. Furthermore, balancing the visual requirements of the study with performance necessitated significant time dedicated to the sourcing, modification, and optimisation of 3D assets to ensure they were "engine-ready". Challenges encountered included complex asset acquisition, as models often required extensive preprocessing (e.g., polygon reduction and texture mapping fixes). Technological challenges arose due to the maturing state of the state-of-the-art Varjo XR-3 hardware and its associated software plug-ins, resulting in inconsistent behaviour and lengthy troubleshooting. The development process was also iterative, driven by feedback from the interdisciplinary research team, which occasionally led to communication gaps, mismatched expectations, and significant unplanned work. Despite these hurdles, the project provides valuable opportunities, establishing a replicable framework for VR cycling simulations and reusable core components. The experience refined optimisation techniques and generated an open-source template that offers learning experiences for future multi-disciplinary research collaborations. The GREENTRAVEL VR project establishes a replicable framework for cycling simulations, offering valuable insights and reusable components for future research leveraging immersive environments in urban street intervention studies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.