CUDA-Accelerated Minimum Distance Query
Minimum distance with gradient among massive number of meshes using CUDA
This project presents a parallel approach to measure the closest distance between 3D convex shapes, an important task in computer graphics and robotics. The CUDA implementation of GJK algorithm achieved a 20x speedup over the CPU implementation with the same scene.
Real-World Application
This project is the default minimum distance query implementation in CFSTrajOpt.jl, where it enables real-time collision avoidance and motion planning by efficiently computing distances between robot links and environmental obstacles.
Overview of the CUDA-accelerated minimum distance query implementation.
Technical Report
For more detailed information about the implementation and performance analysis, please refer to the technical report: