Investigation of the Suitability of 3D Gaussian Splatting for Generating Static Environments for Photorealistic Real-Time Simulations

Context

The ongoing development of highly automated driving functions requires innovative approaches to simulation and validation. Photorealistic real-time simulations make it possible to recreate the traffic environment in detail and test the interaction between automated vehicles and their environment under various conditions. Here, the 3D Gaussian Splatting approach—derived from the radiance field division—offers a new perspective. With its ability to enable high-quality rendering in real time, this approach represents a promising way to bring the real world into a simulation environment. Applying this method to generate static environments has the potential to significantly accelerate and improve the development and testing of highly automated driving functions. In particular, the testing of AI-based or camera-based driving functions can benefit from this method. 

Targets

• Generation of static environments using 3D Gaussian splatting.
• Investigation of the challenges involved in rendering dynamic objects.
• Development of filtering methods for dynamic objects.
• Comparison of quality—particularly in terms of photorealism—with existing assets from game engines. Suitable metrics will be developed for this purpose.
• Investigation of the scalability of 3D Gaussian Splatting with respect to the size of datasets or models.

Prerequisites

• Basic knowledge of computer graphics and rendering.
• Knowledge of Blender and Unreal Engine 5 is a plus
• Programming skills, preferably in Python or C++.
• Willingness to learn advanced rendering techniques.
• Ability to work and think independently