Raymobtime is a methodology for generating realistic wireless communication datasets using ray-tracing combined with 3D dynamic scenarios. The framework ensures physical consistency across time, frequency, and space, enabling high-fidelity modeling of mobility and channel evolution. To enrich the datasets for machine learning, sensing, and multimodal research, we incorporate: LiDAR simulations (via Blensor), RGB camera rendering (via Blender), Position and mobility metadata, Ray-tracing (computed with Remcom Wireless InSite), Vehicle and pedestrian mobility (modeled by SUMO) and Realistic urban maps from CadMapper and OpenStreetMap.
Raymobtime datasets have been used in research on beam selection, beam tracking, blockage prediction, V2V modeling, and multimodal AI for next-generation mmWave/THz communications. For technical details, please refer to our publications.