Real-Time Triangle-SDF Continuous Collision Detection

This document is a research article that presents a novel method for continuous collision detection (CCD) between triangle geometry and signed distance fields (SDFs). The authors introduce a spatio-temporal local optimization approach to determine the time of impact between a triangle and an SDF isosurface. This method aims to improve robustness compared to point sampling techniques and outperforms existing triangle-SDF discrete collision detection algorithms. Additionally, the article discusses a technique for adaptively refining potential collision points on large triangles, allowing for efficient simulations with reduced geometry. The authors evaluate their method's performance in terms of speed, accuracy, and robustness through various complex collision scenarios involving triangles and SDFs. The results indicate that the proposed algorithm is suitable for real-time applications, achieving interactive speeds even in single-threaded CPU implementations. The research addresses significant challenges in collision detection, particularly for fast-moving objects and complex geometries.