The RF scene is represented by Gaussian primitives with mean μ, covariance Σ, and complex-valued radiance ψ and transmittance ρ, whose attributes are updated via gradient-based optimization with adaptive density control. For rendering, rays γ are emitted from the receiver, Gaussians are splatted onto a 2D receiving RF plane, and the received data is obtained by aggregating complex-valued contributions along each ray.
Synthesizing radio-frequency (RF) data given the transmitter and receiver positions, e.g., received signal strength indicator (RSSI), is critical for wireless networking and sensing applications, such as indoor localization. However, it remains challenging due to complex propagation interactions, including reflection, diffraction, and scattering. State-of-the-art neural radiance field (NeRF)-based methods achieve high-fidelity RF data synthesis but are limited by long training times and high inference latency. We introduce GSRF, a framework that extends 3D Gaussian Splatting (3DGS) from the optical domain to the RF domain, enabling efficient RF data synthesis. GSRF realizes this adaptation through three key innovations: First, it introduces complex-valued 3D Gaussians with a hybrid Fourier–Legendre basis to model directional and phase-dependent radiance. Second, it employs orthographic splatting for efficient ray–Gaussian intersection identification. Third, it incorporates a complex-valued ray tracing algorithm, executed on RF-customized CUDA kernels and grounded in wavefront propagation principles, to synthesize RF data in real time. Evaluated across various RF technologies, GSRF preserves high-fidelity RF data synthesis while achieving significant improvements in training efficiency, shorter training time, and reduced inference latency.
Given a transmitter sending RF signals at location (xtx, ytx, ztx), the goal is to synthesize the spatial spectrum received by the receiver (equipped with an antenna array). The spatial spectrum, represented as a 360 × 90 matrix, captures the signal power from all directions around the receiver, covering azimuth and elevation angles at a one-degree resolution.
Given the sparse training dataset setting, GSRF outperforms other methods in both training/inference efficiency and synthesis quality.
@inproceedings{Yang2025_GSRF,
author = {Kang Yang and Gaofeng Dong and Sijie Ji and Wan Du and Mani Srivastava},
title = {GSRF: Complex-Valued 3D Gaussian Splatting for Efficient Radio-Frequency Data Synthesis},
booktitle = {Proceedings of the 39th Conference on Neural Information Processing Systems (NeurIPS)},
year = {2025},
}