Next-generation mobile communication systems are planned to support millimeter Wave (mmWave) transmission in scenarios with high-mobility, such as in private industrial networks. To cope with propagation environments with unprecedented challenges, data-driven methodologies such as Machine Learning (ML) are expected to act as a fundamental tool for decision support in future mobile systems. However, high-quality measurement datasets need to be made available to the research community in order to develop and benchmark ML-based methodologies for next-generation wireless networks. We present a reliable testbed for collecting channel measurements at sub-6 GHz and mmWave frequencies. Further, we describe a rich dataset collected using the presented testbed. Our public dataset enables the development and testing of innovative ML-based channel simulators for both sub-6GHz and mmWave bands on real-world data. We conclude this paper by discussing promising experimental results on two illustrative ML tasks leveraging on our dataset, namely, channel impulse response forecasting and synthetic channel transfer function generation, upon which we propose future exploratory research directions. The original dataset employed in this work is available on IEEE DataPort (https://dx.doi.org/10.21227/3tpp-j394), and the code utilized in our numerical experiments is publicly accessible via CodeOcean (https://codeocean.com/capsule/9619772/tree).

Multi-Band Measurements for Deep Learning-Based Dynamic Channel Prediction and Simulation

Di Cicco, N;Tornatore, M;
2023-01-01

Abstract

Next-generation mobile communication systems are planned to support millimeter Wave (mmWave) transmission in scenarios with high-mobility, such as in private industrial networks. To cope with propagation environments with unprecedented challenges, data-driven methodologies such as Machine Learning (ML) are expected to act as a fundamental tool for decision support in future mobile systems. However, high-quality measurement datasets need to be made available to the research community in order to develop and benchmark ML-based methodologies for next-generation wireless networks. We present a reliable testbed for collecting channel measurements at sub-6 GHz and mmWave frequencies. Further, we describe a rich dataset collected using the presented testbed. Our public dataset enables the development and testing of innovative ML-based channel simulators for both sub-6GHz and mmWave bands on real-world data. We conclude this paper by discussing promising experimental results on two illustrative ML tasks leveraging on our dataset, namely, channel impulse response forecasting and synthetic channel transfer function generation, upon which we propose future exploratory research directions. The original dataset employed in this work is available on IEEE DataPort (https://dx.doi.org/10.21227/3tpp-j394), and the code utilized in our numerical experiments is publicly accessible via CodeOcean (https://codeocean.com/capsule/9619772/tree).
2023
Machine Learning
Dynamic channel
Wireless networks
Channel prediction
Channel simulation
Millimeter wave communications
File in questo prodotto:
File Dimensione Formato  
_ComMag_2023____Multi_band_Measurements_for_Deep_Learning_based_Dynamic_Channel_Prediction_and_Simulation.pdf

accesso aperto

Descrizione: Author-accepted manuscript
: Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione 8.32 MB
Formato Adobe PDF
8.32 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11311/1255397
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact