Deep Learning Aided Diffracting Sphere Radius Determination for Indoor Ray Tracing Algorithm

To address the issue of significant power errors at the receiving end caused by uneven diffraction counts of multiple rays after diffraction, a neural network combining ray tracing algorithms with deep learning is proposed to predict the optimal diffraction sphere radius. This model predicts the appropriate diffraction sphere radius to ensure that multiple rays undergo uniform diffraction counts when encountering objects, thereby improving the accuracy of ray diffraction.