A dual-branch cross-attention neural network with recurrent tracking reconstructs complete channel impulse responses from satellite imagery by predicting TDL parameters, reaching over 0.96 PDP cosine similarity on unseen sites.
Structural 3D reconstruction of indoor space for 5G signal simulation with mobile laser scanning point clouds
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.IV 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Deep Learning-Based Site-Specific Channel Modeling and Inference
A dual-branch cross-attention neural network with recurrent tracking reconstructs complete channel impulse responses from satellite imagery by predicting TDL parameters, reaching over 0.96 PDP cosine similarity on unseen sites.