Modelo de Predição de Perda de Propagação do Sinal de Ondas Sub-6GHz Utilizando Aprendizado Profundo e Dados Visuais Geoespaciais
Resumo
O planejamento eficiente de radiocomunicações, especialmente em frequências sub-6GHz, exige estimativas precisas de perda de propagação. Este trabalho propõe uma arquitetura de Aprendizado Profundo que integra dados numéricos e visuais geoespaciais 2.5D, aplicando o mecanismo de Atenção Parcial para ponderar a relação de obstáculos ao longo do trajeto do sinal. O modelo é treinado e validado com dados reais de medição em cidades alemãs, atingindo um erro quadrático médio (RMSE) de 1,97 dB e Erro Médio Absoluto (MAE) de 1,18 dB, que supera as abordagens tradicionais. Adicionalmente, a transferência de aprendizado na cidade de Moers para Frankfurt demonstra a robustez do método, com RMSE de 2,41 e MAE de 1,89 dB.Referências
Cavalcanti, B., Cavalcante, G., and Mendonça, L. (2025). Performance analysis of artificial neural networks for predicting propagation losses in suburban environments for 4g lte and 5g networks. Revista Principia, 62.
Demir, I., Koperski, K., Lindenbaum, D., Pang, G., Huang, J., Basu, S., Hughes, F., Tuia, D., and Raskar, R. (2018). Deepglobe 2018: A challenge to parse the earth through satellite images. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops.
Lee, J.-Y., Kang, M. Y., and Kim, S.-C. (2019). Path loss exponent prediction for outdoor millimeter wave channels through deep learning. In 2019 IEEE Wireless Communications and Networking Conference (WCNC), pages 1–5.
Park, D., Ryu, S., Lee, S., Kang, N., Kim, S., Kim, K., Choi, D., and Ahn, S. (2024). 5g base station electromagnetic field strength estimation method in complex hotspot area using deep learning. In 2024 IEEE International Symposium on Electromagnetic Compatibility, Signal & Power Integrity (EMC+SIPI), pages 551–554.
Sani, U. S., Malik, O. A., and Lai, D. T. C. (2022). Improving path loss prediction using environmental feature extraction from satellite images: Hand-crafted vs. convolutional neural network. Applied Sciences, 12:7685.
Vasudevan, M. and Yuksel, M. (2024). Machine learning for radio propagation modeling: A comprehensive survey. IEEE Open Journal of the Communications Society.
Waheed, M. T., Fahmy, Y., and Khattab, A. (2022a). Deepchannel: Robust multi-modal outdoor channel model prediction in lte networks using deep learning. IEEE Access, 10:79289–79300.
Waheed, M. T., Khattab, A., and Fahmy, Y. (2022b). Highly accurate multi-modal lte channel prediction via semantic segmentation of satellite images. In 2022 10th International Japan-Africa Conference on Electronics, Communications, and Computations (JAC-ECC), pages 90–93.
Wang, A., Chen, H., Liu, L., Chen, K., Lin, Z., Han, J., and Ding, G. (2024). Yolov10: Real-time end-to-end object detection.
Yeaser, K. M. A. and Hassan, K. M. A. (2025). A comprehensive study on path loss estimation using deep hybrid learning in 5G networks. Journal of Telecommunications and Information Technology, pages 86–94.
Zeng, S., Ji, Y., Chen, W., Yan, L., and Zhao, X. (2024). Mobile network coverage prediction using multi-modal model based on deep neural networks and semantic segmentation. Sensors, 24(16).
Demir, I., Koperski, K., Lindenbaum, D., Pang, G., Huang, J., Basu, S., Hughes, F., Tuia, D., and Raskar, R. (2018). Deepglobe 2018: A challenge to parse the earth through satellite images. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops.
Lee, J.-Y., Kang, M. Y., and Kim, S.-C. (2019). Path loss exponent prediction for outdoor millimeter wave channels through deep learning. In 2019 IEEE Wireless Communications and Networking Conference (WCNC), pages 1–5.
Park, D., Ryu, S., Lee, S., Kang, N., Kim, S., Kim, K., Choi, D., and Ahn, S. (2024). 5g base station electromagnetic field strength estimation method in complex hotspot area using deep learning. In 2024 IEEE International Symposium on Electromagnetic Compatibility, Signal & Power Integrity (EMC+SIPI), pages 551–554.
Sani, U. S., Malik, O. A., and Lai, D. T. C. (2022). Improving path loss prediction using environmental feature extraction from satellite images: Hand-crafted vs. convolutional neural network. Applied Sciences, 12:7685.
Vasudevan, M. and Yuksel, M. (2024). Machine learning for radio propagation modeling: A comprehensive survey. IEEE Open Journal of the Communications Society.
Waheed, M. T., Fahmy, Y., and Khattab, A. (2022a). Deepchannel: Robust multi-modal outdoor channel model prediction in lte networks using deep learning. IEEE Access, 10:79289–79300.
Waheed, M. T., Khattab, A., and Fahmy, Y. (2022b). Highly accurate multi-modal lte channel prediction via semantic segmentation of satellite images. In 2022 10th International Japan-Africa Conference on Electronics, Communications, and Computations (JAC-ECC), pages 90–93.
Wang, A., Chen, H., Liu, L., Chen, K., Lin, Z., Han, J., and Ding, G. (2024). Yolov10: Real-time end-to-end object detection.
Yeaser, K. M. A. and Hassan, K. M. A. (2025). A comprehensive study on path loss estimation using deep hybrid learning in 5G networks. Journal of Telecommunications and Information Technology, pages 86–94.
Zeng, S., Ji, Y., Chen, W., Yan, L., and Zhao, X. (2024). Mobile network coverage prediction using multi-modal model based on deep neural networks and semantic segmentation. Sensors, 24(16).
Publicado
25/05/2026
Como Citar
GALDINO, Caio P.; CAMINHA, Pedro Henrique Cruz; COUTO, Rodrigo de Souza.
Modelo de Predição de Perda de Propagação do Sinal de Ondas Sub-6GHz Utilizando Aprendizado Profundo e Dados Visuais Geoespaciais. In: SIMPÓSIO BRASILEIRO DE REDES DE COMPUTADORES E SISTEMAS DISTRIBUÍDOS (SBRC), 44. , 2026, Praia do Forte/BA.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2026
.
p. 1010-1023.
ISSN 2177-9384.
DOI: https://doi.org/10.5753/sbrc.2026.19335.
