Temporal–Spatial Wavefront Sequence Dataset

This dataset accompanies the study "Maxwell-Informed Deep Network for Spatial Wireless Channel Modeling". Code: https://github.com/WangShuchen97/EMPhyNet The datasets are: 1. matlab.zipThis package provides both the scenario data and the generated Temporal–Spatial Wavefront Sequence (TSWS) dataset. The folder ./matlab/map_data contains environmental maps derived from OpenStreetMap, which are utilized in the Ray Tracing process implemented in ./matlab/main.m to produce the TSWS dataset ./matlab/data. The file also includes the processed building data, stored in ./matlab/map_height. 2. data.zipThis package provides datasets for training and testing the EMPhyNet model, compiled by the script ./data_m2pt.py using the TSPFS data (./matlab/data) and building distribution information (./matlab/map_height). 3. checkpoints.zipThis package includes the pretrained EMPhyNet model parameters, comprising three specialized submodels responsible for wavefront trajectory prediction, real component estimation, and imaginary component estimation, respectively. These checkpoints can be directly loaded for evaluation or fine-tuning. 4. results.zipThis file contains the simulation and inference results generated by the trained models. The results can be visualized and evaluated using the script Visualization_and_MSE.py.