On the Limitations of Ray-Tracing for Learning-Based RF Tasks in Urban Environments
A new research paper explores the limitations of ray-tracing simulations for learning-based radio frequency (RF) tasks in urban environments. The study, conducted in Rome, found that while precise geometry and antenna models are crucial for simulation fidelity, accurately capturing residual urban noise remains a significant challenge for transferable RF simulations. The research highlights that antenna locations and orientations have a decisive impact on simulation accuracy, with simple optimization improving correlation by up to 130% and reducing localization error by one-third. AI
