Collision detection is a fundamental problem in computer graphics, physical simulation, and robotics, and numerous algorithms have been proposed for different object representations and application scenarios. Signed distance fields (SDFs) offer particular advantages for complex geometries: they provide O(1)-time distance queries from arbitrary points in space to the object surface and naturally support the computation of penetration depth, contact points, and contact normal. As an implicit representation that encodes geometry as a scalar field, SDFs can be realized using analytic functions, voxel grids, or neural network–based models, and are often more suitable than traditional mesh-based methods for handling ob-jects with complex continuous surfaces.
The objective of this thesis is to apply SDF-based collision detection methods to an exper-imental digital twin platform and to systematically assess their strengths and limitations in complex simulation scenarios. The work will comprise three main parts: a literature review of state-of-the-art collision detection algorithms and their application domains; the imple-mentation of an SDF-based collision detection method in a C++ simulation platform with GPU acceleration for large-scale parallel computation; and a quantitative benchmark against other mainstream collision detection approaches on representative test cases in terms of performance and accuracy.
Keywords: Physics-based simulation, game engine, collision detection, GPU
Requirements:
- You are studying Electrical Engineering, Automation, Robotics, Computer Science, or a related field.
- You have a strong interest in game physics engines or robot simulation; ideally, you have studied multibody dynamics or robotics dynamics.
- Ideally, you have programming skills in C++.
Relevant Literature:
- Liu, P., Zhang, Y., Wang, H., Yip, M. K., Liu, E. S., & Jin, X. (2024). Real-time col-lision detection between general SDFs. Computer Aided Geometric Design, 111, 102305.
- Macklin, M., Erleben, K., Müller, M., Chentanez, N., Jeschke, S., & Corse, Z. (2020). Local optimization for robust signed distance field collision. Proceedings of the ACM on Computer Graphics and Interactive Techniques, 3(1), 1-17.
Betreuer: Shao, Email:
