EA-RMENet -- Path Loss Prediction in Urban Environments using Deep Learning
Summary
Proposes EA-RMENet, a deep learning model for radio map estimation that achieves high accuracy and efficiency using an EfficientNetB5 encoder, attention gated skip connections, and ASPP.
View Cached Full Text
Cached at: 07/21/26, 06:48 AM
# EA-RMENet -- Path Loss Prediction in Urban Environments using Deep Learning Source: [https://arxiv.org/abs/2607.16449](https://arxiv.org/abs/2607.16449) Authors:[Jonathan O'Shea](https://arxiv.org/search/cs?searchtype=author&query=O%27Shea,+J)\(DCU School of Electronic Engineering\),[Conor Brennan](https://arxiv.org/search/cs?searchtype=author&query=Brennan,+C)\(DCU School of Electronic Engineering\) [View PDF](https://arxiv.org/pdf/2607.16449) > Abstract:Accurate path loss prediction is a critical component of wireless network planning\. Current path loss prediction methods typically struggle to balance the trade\-off between accuracy and computational efficiency\. This paper proposes the Efficient Attention Radio Map Estimation Network \(EA\-RMENet\) which is an image data\-driven, deep learning \(DL\) model designed for radio map estimation \(RME\)\. EA\-RMENet uses a U\-Net framework with an EfficientNetB5 encoder, Attention Gated \(AG\) skip connections, and Atrous Spatial Pyramid Pooling \(ASPP\)\. The EfficientNet encoder uses compound scaling to balance accuracy and efficiency\. AG skip connections suppress irrelevant features, and the ASPP captures a multi\-scale context\. The model has a test prediction RMSE of 0\.0334 on the RadioMapSeer3D dataset with an inference time of 0\.022 seconds/sample\. In the ICASSP 2023 Radio Map Prediction Challenge, the model ranks third with a competitive RMSE of 0\.0406 this highlights the models potential for real\-world RME\. ## Submission history From: Jonathan O'Shea O'Shea \[[view email](https://arxiv.org/show-email/0d9afaa5/2607.16449)\] **\[v1\]**Fri, 17 Jul 2026 18:47:46 UTC \(917 KB\)
Similar Articles
PDRNN: Modular Data-driven Pedestrian Dead Reckoning on Loosely Coupled Radio- and Inertial-Signalstreams
Proposes PDRNN, a modular hybrid AI-assisted pedestrian dead reckoning system that combines a recurrent neural network with separate ML models for orientation, velocity, and distance estimation, with optional radio-based stabilization. Experiments on dynamic sports movement data show superior accuracy and precision compared to classic and ML-based methods.
Neural Radiated-Noise Fields for Unmanned Underwater Vehicle Noise Spectrum Prediction in Three-Dimensional Scenes
This paper proposes Neural Radiated-Noise Fields (NRNF), a neural network approach for predicting underwater vehicle radiated noise spectra as a continuous function of 3D position, orientation, and frequency. Evaluated on lake trial data, the model achieves an average prediction error of 3.5 dB in the 50–5000 Hz band across multiple generalization settings.
MARCUS: Missing-Aware Region Representation with Contextual Urban Signals for Rent Prediction
The paper proposes MARCUS, a missing-aware region representation model that treats missing data as contextual urban signals for rent prediction, achieving state-of-the-art performance with significant error reduction on real-world datasets.
Beyond the Previous Layer: Residual Predictive Structure in Sparse MoE Routing
This paper investigates the predictive value of earlier expert selections in sparse mixture-of-experts models beyond the most recent layer, finding significant gains in routing prediction across layers.
Neural Network-Assisted CLEAN for Channel Modeling in Low-SNR Regimes
This paper proposes NN-CLEAN, a hybrid framework that embeds a multi-head residual network into the iterative CLEAN extraction loop for efficient multipath parameter estimation in low-SNR channel modeling. It matches traditional Grid-Search CLEAN accuracy while greatly reducing computational complexity and enabling parallelization for real-time MIMO systems.