Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Practical Implementation of RIS-Aided Spectrum Sensing: A Deep Learning-Based Solution

Sefa Kayraklik, Ibrahim Yildirim, Ertugrul Basar, Ibrahim Hokelek, Ali Gorcin

TL;DR

This work tackles spectrum sensing for cognitive radio under spectrum scarcity by integrating a reconfigurable intelligent surface (RIS) with deep learning (DL). It converts RIS-aided received signals into spectrogram images and applies two state-of-the-art object detectors, Detectron2 and YOLOv7, to identify the PT signal type and its spectral usage. The approach is validated through both synthesized datasets (LTE/NR signals) and real-world measurements using a RIS prototype, showing substantial gains in detection accuracy and spectral localization when the RIS is optimized. The findings demonstrate the practicality of RIS-enabled DL spectrum sensing for more efficient dynamic spectrum utilization in next-generation wireless networks.

Abstract

This paper presents reconfigurable intelligent surface (RIS)-aided deep learning (DL)-based spectrum sensing for next-generation cognitive radios. To that end, the secondary user (SU) monitors the primary transmitter (PT) signal, where the RIS plays a pivotal role in increasing the strength of the PT signal at the SU. The spectrograms of the synthesized dataset, including the 4G LTE and 5G NR signals, are mapped to images utilized for training the state-of-art object detection approaches, namely Detectron2 and YOLOv7. By conducting extensive experiments using a real RIS prototype, we demonstrate that the RIS can consistently and significantly improve the performance of the DL detectors to identify the PT signal type along with its time and frequency utilization. This study also paves the way for optimizing spectrum utilization through RIS-assisted CR application in next-generation wireless communication systems.

Practical Implementation of RIS-Aided Spectrum Sensing: A Deep Learning-Based Solution

TL;DR

This work tackles spectrum sensing for cognitive radio under spectrum scarcity by integrating a reconfigurable intelligent surface (RIS) with deep learning (DL). It converts RIS-aided received signals into spectrogram images and applies two state-of-the-art object detectors, Detectron2 and YOLOv7, to identify the PT signal type and its spectral usage. The approach is validated through both synthesized datasets (LTE/NR signals) and real-world measurements using a RIS prototype, showing substantial gains in detection accuracy and spectral localization when the RIS is optimized. The findings demonstrate the practicality of RIS-enabled DL spectrum sensing for more efficient dynamic spectrum utilization in next-generation wireless networks.

Abstract

This paper presents reconfigurable intelligent surface (RIS)-aided deep learning (DL)-based spectrum sensing for next-generation cognitive radios. To that end, the secondary user (SU) monitors the primary transmitter (PT) signal, where the RIS plays a pivotal role in increasing the strength of the PT signal at the SU. The spectrograms of the synthesized dataset, including the 4G LTE and 5G NR signals, are mapped to images utilized for training the state-of-art object detection approaches, namely Detectron2 and YOLOv7. By conducting extensive experiments using a real RIS prototype, we demonstrate that the RIS can consistently and significantly improve the performance of the DL detectors to identify the PT signal type along with its time and frequency utilization. This study also paves the way for optimizing spectrum utilization through RIS-assisted CR application in next-generation wireless communication systems.
Paper Structure (9 sections, 4 equations, 9 figures, 1 table)

This paper contains 9 sections, 4 equations, 9 figures, 1 table.

Table of Contents

  1. Introduction
  2. System Model
  3. DL-based Spectrum Sensing
  4. State-of-the-Art Object Detectors: Detectron2 and YOLOv7
  5. Training Parameters of the Detectors
  6. Signals of Interest, Dataset Representation and Generation
  7. Performance Analysis of the Detectors in Practical Measurement
  8. Conclusion
  9. Acknowledgment

Figures (9)

  • Figure 1: The overview of the RIS-empowered spectrum sensing system with an DL-based solution.
  • Figure 2: The workflow of the phases of training and system operation of detectors for DL-based spectrum sensing.
  • Figure 3: The spectrograms of the generated signals of (a) 4G LTE, (b) 5G NR, and (c) both 4G LTE and 5G NR, where the x-axis and y-axis represent the frequency and time domains, respectively.
  • Figure 4: Detectron2's (a) accuracy, (b) total loss, and (c) average precision with the test dataset; YOLOv7's (d) mAP@0.5:0.95, (e) objectness loss, and (f) average precision with the test dataset.
  • Figure 5: The RIS-aided spectrum sensing measurement setup.
  • ...and 4 more figures