IPP-Net: A Generalizable Deep Neural Network Model for Indoor Pathloss Radio Map Prediction
This work addresses indoor radio propagation prediction for wireless communication systems, but it is incremental as it builds on existing UNet architectures and benchmark datasets.
The authors tackled indoor pathloss radio map prediction by proposing IPP-Net, a UNet-based deep neural network model trained on ray tracing simulation and a modified 3GPP model, achieving a weighted root mean square error of 9.501 dB and second place in the ICASSP 2025 challenge.
In this paper, we propose a generalizable deep neural network model for indoor pathloss radio map prediction (termed as IPP-Net). IPP-Net is based on a UNet architecture and learned from both large-scale ray tracing simulation data and a modified 3GPP indoor hotspot model. The performance of IPP-Net is evaluated in the First Indoor Pathloss Radio Map Prediction Challenge in ICASSP 2025. The evaluation results show that IPP-Net achieves a weighted root mean square error of 9.501 dB on three competition tasks and obtains the second overall ranking.