Envelope Level Crossing Rate and Average Fade Duration of a Generic 3D Non-Stationary UAV Channel Model

Qiuming Zhu, Neng Cheng, Xiaomin Chen, Weizhi Zhong, Boyu Hua, Yawen WangIEEE Access 2020 · DOI

Introduction

The rapid proliferation of Unmanned Aerial Vehicles (UAVs) in applications ranging from search and rescue to 5G aerial relays has necessitated the development of highly accurate communication models. A primary challenge in these environments is "fading"—the rapid fluctuation of signal strength caused by multipath interference and obstacles. While fading is well-understood for stationary or 2D-moving ground devices, UAVs present a unique challenge: they move at high speeds in three-dimensional space, often with time-varying velocities and arbitrary trajectories.

Most traditional channel models assume "stationarity," where the statistical properties of the signal remain constant over time. However, the dynamic nature of UAV flight makes the channel inherently non-stationary. This paper introduces a generic 3D non-stationary Geometry-Based Stochastic Model (GBSM) and derives closed-form expressions for critical fading metrics: the Level Crossing Rate (LCR) and Average Fade Duration (AFD). These metrics are essential for engineers designing robust error-correction and interleaving schemes for high-speed aerial links.

Modeling 3D Non-Stationary UAV Channels

To capture the complexity of the UAV communication environment, the researchers utilized a GBSM framework. In this model, the transmitter (UAV) and receiver (e.g., a ground vehicle) are modeled within a 3D space populated by scattering clusters. These clusters, representing physical obstacles like buildings or trees, are modeled as virtual cylinders.

Figure 3: The proposed 3D non-stationary GBSM for UAV-to-vehicle communication, illustrating the geometry of scattering clusters and time-varying trajectories.

The model's primary innovation lies in its support for arbitrary trajectories. Unlike previous models that often simplified movement to straight lines or constant speeds, this model incorporates time-varying 3D velocity vectors for both the transmitter and receiver. This allows for the simulation of complex flight paths, including acceleration and turns, which significantly impact the Doppler spectrum and fading characteristics.

Technical Contributions

  • Arbitrary Trajectories: Integration of time-varying elevation and azimuth angles alongside velocity vectors.
  • Mathematical Rigor: Derivation of complex joint Probability Density Functions (PDFs) for the fading envelope to calculate LCR and AFD.
  • Physical Consistency: The model ensures that the time-variant Doppler frequency is calculated as the derivative of phase over time, maintaining consistency with physical reality.

Validation and Results

The theoretical model was validated through two primary methods: extensive numerical simulations and comparison with empirical measurement data.

Numerical Simulation

The researchers conducted 1,500 Monte Carlo simulations to verify the derived closed-form equations. The results showed that the theoretical LCR and AFD curves matched the simulated data almost perfectly across various time intervals (e.g., $t = 0.5s, 1.0s, 1.5s$). This confirms the mathematical robustness of the model in predicting how often a signal drops below a threshold (LCR) and how long it remains in that faded state (AFD).

Empirical Alignment

The model was further tested against real-world measurement data from urban UAV channel sounding tests conducted at 2.4 GHz. By adjusting the model parameters to match the experimental setup (e.g., a UAV speed of 8 m/s at a distance of 1000m), the authors demonstrated that their model’s LCR and AFD predictions closely followed observed physical reality.

Key Insights for System Design

The study reveals several critical insights for the design of next-generation UAV communication systems:

  • Dynamic Fading Tempo: LCR and AFD are not static; they fluctuate as the UAV moves. A coding scheme optimized for one point in a flight path may become ineffective seconds later as the fading "tempo" changes.
  • Trajectory-Aware Communication: The specific 3D flight path directly dictates the Doppler spectrum. This suggests that future communication systems should be "cross-layer" aware, potentially adjusting transmission parameters based on flight controller data.
  • Geometry Impacts: The height and radius of scattering clusters (the virtual cylinders) significantly influence fading severity. Higher clusters generally lead to more frequent and intense signal fluctuations.
  • Threshold Sensitivity: As the signal threshold increases, the LCR decreases while the AFD increases. This indicates that while the signal may dip into deep fades less frequently, it tends to linger in moderately weak states for longer periods, which is a vital consideration for determining optimal data packet lengths.

Conclusion and Future Directions

This work provides a powerful, generic framework for understanding and predicting fading in 3D non-stationary UAV channels. By providing closed-form expressions for LCR and AFD, it offers a practical tool for engineers to enhance the reliability of aerial wireless links. Future research will focus on expanding the model to more diverse environments, such as mountainous or maritime regions, where scattering geometries differ from the urban "cylinder" model.

Summary updated

BibTeX
@article{zhu2020envelope,
  author = {Qiuming Zhu and Neng Cheng and Xiaomin Chen and Weizhi Zhong and Boyu Hua and Yawen Wang},
  title = {{Envelope Level Crossing Rate and Average Fade Duration of a Generic 3D Non-Stationary UAV Channel Model}},
  journal = {IEEE Access},
  year = {2020},
  doi = {10.1109/access.2020.3014303},
}
Figure