Samenvatting

This thesis aims to answer the question ‘How to create a well-optimized, photorealistic Virtual Reality learning environment in Unity to instruct users about radar maintenance, in which they feel relaxed and comfortable?’ for the client Thales. This international military contractor company is the largest supplier of radar systems in the Netherlands. Thales wanted a new VR learning environment for radar maintenance. The main problem was that a previous VR learning environment was too artificial and discomforting and had performance issues. The proposed solution was to research and create a photorealistic and relaxing prototype that runs well on the target platform and fits Thales’s company profile. Research and development were structured based on the ‘Design Thinking Model’. Online desk research was used to gather already available data. Previous experience in 3D modeling, texturing, sculpting, rendering was utilized. Weekly meetings with Thales provided a constant flow of feedback and enabled iterative development. The testing methods included technical performance tests to inspect optimization, and playtesting and interviews for feedback. Based on online desk research, client discussions, and player feedback, a bright, serene VR learning environment was created. It was set in a lagoon surrounded by green nature scenery. The playable area consisted of an observatory and a platform on which the maintenance would take place. Feedback gathered during playtest sessions aligned with the findings, and testers reported that the scene had an overall relaxing atmosphere. Though photorealism proved to be too ambitious for the size of the team, Thales was satisfied with the final visuals. Since the target platform was VR, it was crucial to have stable frame rates of at least 90. Otherwise, the players may have experienced discomfort and nausea. The key aspects of optimization were the number of triangles, draw calls, and render settings. Optimizing for VR proved to be challenging and time-consuming. Overall, optimization was successful, and stable 90 FPS was reached in the final build. The clients were content with these results.