A Refined Path Generation Pipeline for Radio Channel Propagation Modeling

Ray tracing is a widely used approach for deterministic modelling of radio channels. We present our path generation pipeline, which combines environment discretization-based propagation path search with path refinement, which outputs validated paths fulfilling the Fermat’s principle of the least time. We propose a novel gradient descent-based solution for refinement. Whole pipeline is implemented as GPU computations using NVIDIA CUDA and OptiX ray tracing engine, and experimental results show efficacy of the approach.