EM DeepRay: An Expedient, Generalizable, and Realistic Data-Driven Indoor Propagation Model

Efficient and realistic indoor radio propagation modeling tools are inextricably intertwined with the design operation of next-generation wireless networks. Machine-learning (ML)-based models can be trained simulated or real-world data to provide accurate estimates channel characteristics in a computationally efficient way. However, most existing research works on ML-based focus outdoor modeling, while data-driven remain site-specific limited scalability. In this article, we present an credible framework for environments. Specifically, demonstrate how convolutional encoder–decoder replicate results ray tracer, by encoding physics-based information environment, such as permittivity walls, decoding it path loss (PL) heatmap environment interest. Our model is over multiple geometries frequency bands, eventually predict PL unknown bands within few milliseconds. addition, illustrate concept transfer learning leveraged calibrate our adjusting its preestimate weights, allowing make predictions that consistent measurement data.