Neural Beam Field for Spatial Beam RSRP Prediction

TL;DR

This paper introduces Neural Beam Field (NBF), a hybrid neural-physical framework that accurately predicts spatial beam RSRP in dense wireless networks by combining a learnable physical environment model with neural networks, improving efficiency and generalization.

Contribution

The paper proposes a novel hybrid neural-physical model with a learnable environment intermediary and a physics-inspired module, enhancing beam RSRP prediction accuracy and generalization in wireless networks.

Findings

NBF outperforms traditional channel knowledge maps and pure neural models in accuracy.

NBF demonstrates improved training efficiency and model compactness.

The framework effectively generalizes across different site environments.

Abstract

Accurately predicting beam-level reference signal received power (RSRP) is essential for beam management in dense multi-user wireless networks, yet challenging due to high measurement overhead and fast channel variations. This paper proposes Neural Beam Field (NBF), a hybrid neural-physical framework for efficient and interpretable spatial beam RSRP prediction. Central to our approach is the introduction of the Multi-path Conditional Power Profile (MCPP), a learnable physical intermediary representing the site-specific propagation environment. This approach decouples the environment from specific antenna/beam configurations, which helps the model learn site-specific multipath features and enhances its generalization capability. We adopt a decoupled ``blackbox-whitebox" design: a Transformer-based deep neural network (DNN) learns the MCPP from sparse user measurements and positions, while a physics-inspired module analytically infers beam RSRP statistics. To improve convergence and adaptivity, we further introduce a Pretrain-and-Calibrate (PaC) strategy that leverages ray-tracing priors for physics-grounded pretraining and then RSRP data for on-site calibration. Extensive simulation results demonstrate that NBF significantly outperforms conventional table-based channel knowledge maps (CKMs) and pure blackbox DNNs in prediction accuracy, training efficiency, and generalization, while maintaining a compact model size. The proposed framework offers a scalable and physically grounded solution for intelligent beam management in next-generation dense wireless networks.