Map-based Channel Modeling and Generation for U2V mmWave Communication

Qiuming Zhu, Kai Mao, Maozhong Song, Xiaomin Chen, Boyu Hua, Weizhi Zhong, Xijuan YeIEEE Trans. Vehicular Technology 2021

TL;DR

This paper introduces a hybrid, map-based channel model for UAV-to-Vehicle (U2V) communications at millimeter-wave (mmWave) frequencies. By combining deterministic ray tracing for large-scale paths with a stochastic Gaussian Mixture Model (GMM) for fine-grained ray clusters, the researchers created a simulation framework that is both computationally efficient and highly accurate to real-world 3D environments.

What Problem Does This Paper Address?

As we move toward 5G-Beyond and 6G, Unmanned Aerial Vehicles (UAVs) are becoming critical for flexible network deployment. However, communicating between a high-speed drone and a ground vehicle using mmWave frequencies is a modeling nightmare.

Standard models often fail because:

  1. Non-stationarity: The channel characteristics change rapidly as the UAV moves through 3D space.
  2. Complexity: mmWave signals behave like light (quasi-optical), meaning every building and tree matters.
  3. Computational Cost: Pure Ray Tracing (RT)—simulating every single photon-like path—is too slow for practical use, while purely statistical models are too generic to reflect specific city layouts.

The authors aim to bridge this gap by creating a "map-based" model that respects the specific geometry of a city while remaining fast enough for researchers to use.

Methods

The authors propose a hybrid framework that splits the channel into two parts:

  • Deterministic Inter-path Modeling: The "main" paths (the big reflections off buildings) are calculated using a simplified 3D reconstruction of digital maps (like those from Google Earth). They convert complex terrain into regular triangle facets to speed up the math.
  • Stochastic Intra-path Modeling: Instead of tracing every tiny sub-ray (which is computationally expensive), they group rays into clusters. They use a Gaussian Mixture Model (GMM) to statistically describe the delay and angle offsets within these clusters.
  • Modified Method of Equal Areas (MMEA): To turn these statistical GMMs into actual random variables for simulation, they developed MMEA, which allows for efficient generation of random intra-path parameters that match the observed data.
  • 3D Factors: The model explicitly accounts for the UAV’s 3D trajectory, its rotation, and the specific orientation of its 3D antenna array, all of which cause complex Doppler shifts and phase changes.

Datasets

The researchers generated a massive dataset using high-fidelity Ray Tracing across four typical scenarios: Urban, Hilly, Forest, and Sea. * RT Data: 25,000 pairs of transceivers were simulated (50 ground receivers and 500 UAV positions) to extract statistical properties. * Validation Data: The model’s results were compared against existing theoretical models and real-world measurement data (specifically citing measurements at 24 GHz from prior literature) to ensure the Autocorrelation Function (ACF) and Power Delay Profiles (PDP) were realistic.

Key Insights and Ideas

  • GMM > Standard Gaussian: The most significant "aha" moment is that standard Gaussian distributions (often used in 3GPP standards) don't actually fit the delay and angle offsets of mmWave rays very well. The GMM (which combines multiple Gaussians) provides a much tighter fit to the "massive" ray-tracing data.
  • Map Simplification is Key: By reconstructing the digital map into simplified triangle facets, the authors reduced the complexity of the environment without losing the essential "quasi-optical" characteristics of mmWave propagation.
  • Birth-and-Death Process: The model successfully captures the "birth-and-death" of signal paths—how a connection might suddenly appear as a drone clears a building and disappear just as quickly as it passes behind another.

Key Results

  • Superior Fitting: The GMM achieved significantly lower Root Mean Square Error (RMSE) compared to traditional models. For example, in the urban delay offset scenario, GMM had an RMSE of 0.0021, while the standard Gaussian Distribution (GD) was much higher at 0.0317.
  • Angle Accuracy: For azimuth offsets, GMM (RMSE 0.0028) vastly outperformed the commonly used Von Mises Distribution (RMSE 0.0301).
  • Physical Consistency: The generated Doppler Power Spectrum Density (DPSD) and Autocorrelation Functions (ACF) showed the same trends as deterministic RT simulations and real-world measurements, proving the hybrid approach doesn't sacrifice physical reality for speed.

Limitations and Future Work

While the model is robust, the authors note that the "intra-path" parameters (the fine details within a cluster) still rely heavily on the initial massive ray-tracing data. They suggest that future work should involve more extensive field measurements to further refine these stochastic parameters, especially for diverse weather conditions or highly dynamic environments with moving obstacles like buses or trains.

Summary updated

Abstract

Unmanned aerial vehicle (UAV) aided millimeter wave (mmWave) technologies have a promising prospect in the future communication networks. By considering the factors of three-dimensional (3D) scattering space, 3D trajectory, and 3D antenna array, a non-stationary channel model for UAV-to-vehicle (U2V) mmWave communications is proposed. The computation and generation methods of channel parameters including interpath and intra-path are analyzed in detail. The inter-path parameters are calculated in a deterministic way, while the parameters of intra-path rays are generated in a stochastic way. The statistical properties are obtained by using a Gaussian mixture model (GMM) on the massive ray tracing (RT) data. Then, a modified method of equal areas (MMEA) is developed to generate the random intra-path variables. Meanwhile, to reduce the complexity of RT method, the 3D propagation space is reconstructed based on the user-defined digital map. The simulated and analyzed results show that the proposed model and generation method can reproduce non-stationary U2V channels in accord with U2V scenarios. The generated statistical properties are consistent with the theoretical and measured ones as well.
BibTeX
@article{zhu2021map,
  author = {Qiuming Zhu and Kai Mao and Maozhong Song and Xiaomin Chen and Boyu Hua and Weizhi Zhong and Xijuan Ye},
  title = {{Map-based Channel Modeling and Generation for U2V mmWave Communication}},
  journal = {IEEE Trans. Vehicular Technology},
  year = {2021},
}
Figure