Multi-Modal Sensing Residual-Corrected GNN for mmWave Path Loss Prediction via Synesthesia of Machines
Mengyuan Lu, Lu Bai, Xiang Cheng · 2026
Read Multi-Modal Sensing Residual-Corrected GNN for mmWave Path Loss Prediction via Synesthesia of Machines by Mengyuan Lu, Lu Bai, Xiang Cheng (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-18005.txt.
Description
Mengyuan Lu and team propose a multi-modal sensing residual-corrected graph neural network for millimeter-wave path loss prediction in vehicular links for future intelligent transport systems. The setting becomes an environment sensing path loss graph, with Tx and Rx nodes, Tx-Rx links, and Rx-Rx spatial links. A geometry-driven baseline splits steady trends from random residuals.
Topology-aware graphs and visual detail join through gated fusion to predict the path loss gap from that base. The team built a path loss dataset spanning an urban wide lane, an urban crossroad, and a suburban forking road, matching camera images and scene data to ray-tracing data. It beat empirical and standard baselines, with normalized mean squared error under one hundredth, mean absolute error under six decibels, and mean absolute percentage error near five percent. Few-shot fine-tuning extended it to unseen places.
Catalog record
| CREATOR | Mengyuan Lu, Lu Bai, Xiang Cheng |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-18005