Transmission-Integrated HAIT Ray Tracing for Massive O2I Propagation Modeling With MTN Preprocessing and Anxel Beam Transformation

Despite its computational efficiency and accuracy, the heterogeneous computing-based anxel beam shrinkage method-accelerated image theory (HAIT) ray tracing (RT) method has been limited to outdoor scenarios due to the lack of transmission modeling. Existing HAIT approaches face challenges in transmission-inclusive environments, demanding substantial computational acceleration. This article proposes a transmission-integrated HAIT (TI-HAIT) RT framework for large-scale outdoor-to-indoor (O2I) propagation modeling. TI-HAIT introduces two key innovations: 1) a GPU-based minimum transmission number preprocessing (MTNP) algorithm for efficient transmission handling and 2) an anxel beam transformation (ABT) method for optimized angular-Z buffer (AZB) matrix utilization. Leveraging central processing unit (CPU)/graphical processing unit (GPU) parallel computing, TI-HAIT achieves up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$6051\times $ </tex-math></inline-formula> speedup over the image theory (IT) solver in WinProp. MTNP accelerates simulations by up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$153\times $ </tex-math></inline-formula> and reduces CPU memory usage by 78%, while ABT improves visibility tree generation by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$4.1\times $ </tex-math></inline-formula> and reduces CPU/GPU memory usage by 7% and 9%, respectively. The framework enables simulations of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.1\times 1.1$ </tex-math></inline-formula> km urban areas with large indoor structures, supporting up to five bounces, four transmissions, and hundreds of field observation points. TI-HAIT advances IT-based RT by enabling efficient and accurate transmission analysis in complex scenarios previously infeasible with conventional IT-based RT methods.