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.
DL- Enhanced channel parameter prediction scheme based on adaptive meta mask R-CNN model
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.