UAV-Assisted Three-Dimensional Spectrum Mapping Driven by Spectrum Data and Channel Model

Xiaofu Du, Qiuming Zhu, Guoru Ding, Jie Li, Qihui Wu, Tianxu Lan, Zhipeng Lin, Weizhi Zhong, Lu HanSymmetry 2021 · DOI

TL;DR

This paper introduces a novel UAV-assisted system for creating detailed 3D spectrum maps, crucial for managing increasingly scarce radio frequencies. By combining spectrum data with a propagation channel model and employing a unique uniformity-driven data completion algorithm, the system can reconstruct accurate maps even with limited and unevenly distributed measurements.

What Problem Does This Paper Address?

Spectrum scarcity and security are growing concerns, extending from terrestrial networks to aerial and space-based communications. Effectively managing and utilizing limited spectrum resources requires detailed knowledge of their distribution. Spectrum maps, which visualize radio parameters like signal strength, are vital for this. However, existing methods often focus on 2D maps or struggle with incomplete and unevenly distributed data, especially when trying to create comprehensive 3D representations needed for integrated communication networks. This paper aims to bridge this gap by developing a robust 3D spectrum mapping system that can overcome these data limitations.

Methods

The core of this work is a UAV-assisted spectrum mapping system that leverages both collected spectrum data and a propagation channel model. The process involves:

  • UAV-based Data Collection: A UAV equipped with both omnidirectional and directional antennas surveys the measurement area to collect initial spectrum data and GPS information.
  • Model-Driven Spectrum Data Inference: Using the directional antenna's data and a propagation channel model (specifically a log-normal shadowing model adapted for UAV heights), the system infers spectrum data for locations within the directional antenna's beam. This step utilizes the antenna's directivity and the channel's physical characteristics.
  • Multi-Source Data Fusion: When multiple inferred or measured data points exist for the same location, a fusion method combines them to determine the most accurate RSS value.
  • Model-Driven Spectrum Data Completion with Uniformity Decision Mechanism: This is the paper's key innovation. After inference, some areas may still have unknown spectrum values. This method first assesses the spatial uniformity of the known (measured and inferred) data in the vicinity of an unknown cube.
  • High-Uniformity Case: If the known data is relatively uniform, an enhanced Inverse Distance Weighted (IDW) method (modified Shepard's method) is used, incorporating distance, angle, and path loss factors.
  • Low-Uniformity Case: If the known data is unevenly distributed, a 3D equivalent radiation source location estimation method (based on the LIvE method) is employed.
  • 3D Spectrum Map Reconstruction: Finally, the completed data is used to construct the full 3D spectrum map.

Datasets

The paper uses simulated data generated via a ray-tracing (RT) technique for evaluation. This RT-based simulation mimics a typical campus scenario with buildings and a single radiation source, providing ground truth data for performance comparison. No new datasets were introduced.

Key Insights and Ideas

  • The Power of Directionality and Channel Models: The paper demonstrates that effectively combining directional antenna data with a physically grounded propagation channel model is crucial for accurate spectrum data inference, especially in complex environments.
  • Uniformity as a Key Decision Factor: The novel approach of using the spatial uniformity of existing data to dynamically select the most appropriate completion method (high-uniformity vs. low-uniformity) is a significant contribution. This adaptive strategy allows for better performance across diverse data distribution scenarios.
  • UAVs as Versatile Spectrum Sensors: The use of UAVs, with their ability to cover large areas and carry sophisticated sensors (directional and omnidirectional antennas), is highlighted as a practical solution for acquiring the necessary multi-dimensional spectrum information.
  • 3D Mapping is Essential for Modern Networks: The paper implicitly argues for the necessity of 3D spectrum maps, which are vital for understanding and managing spectrum in increasingly complex aerial and integrated communication systems.

Key Results

The proposed UAV-assisted spectrum mapping scheme demonstrates superior performance compared to classical IDW and tensor completion methods.

  • Reconstruction Accuracy: The method achieves approximately 12.5% better reconstruction accuracy than IDW and a remarkable 92.3% better accuracy than tensor completion.
  • Performance Under Data Limitations: These improvements are particularly pronounced when the collected spectrum data is regularly missing, unevenly distributed, and limited.
  • Simulation Validation: The results are validated using ray-tracing simulated data of a campus environment, showing the effectiveness of the approach in a realistic scenario.

Limitations and Future Work

  • Limitations:
    • The paper acknowledges that field measurements for UAV channels are "very hard and costly," which is why they rely on RT simulations for evaluation.
  • The current work primarily focuses on Received Signal Strength (RSS).
  • Future Work:
    • The authors plan to incorporate frequency and time information into the spectrum measurement process. This will significantly increase the amount of information captured in the reconstructed multi-dimensional spectrum map, leading to even richer insights.

Summary updated

BibTeX
@article{du2021uav,
  author = {Xiaofu Du and Qiuming Zhu and Guoru Ding and Jie Li and Qihui Wu and Tianxu Lan and Zhipeng Lin and Weizhi Zhong and Lu Han},
  title = {{UAV-Assisted Three-Dimensional Spectrum Mapping Driven by Spectrum Data and Channel Model}},
  journal = {Symmetry},
  year = {2021},
  doi = {10.3390/sym13122308},
}
Figure