Abstract
Extremely large antenna array (ELAA) is key to enhancing spectral efficiency in 6G networks. Leveraging the distributed nature of multi-unmanned aerial vehicle (UAV) systems enables the formation of distributed ELAA, which often operate in the near-field region with spatial sparsity, rendering the conventional far-field plane wave assumption invalid. This paper investigates channel estimation for distributed near-field multi-UAV communication systems. We first derive closed-form signal-to-noise ratio (SNR) expressions under the plane wave model (PWM), spherical wave model (SWM), and a hybrid spherical-plane wave model (HSPWM), also referred to as the cross-field model, within a distributed uniform planar array (UPA) scenario. The analysis shows that HSPWM achieves a good balance between modeling accuracy and analytical tractability. Based on this, we propose two channel estimation algorithms: the spherical-domain orthogonal matching pursuit (SD-OMP) and the tensor-OMP. The SD-OMP generalizes the polar domain to jointly consider elevation, azimuth, and range. Under the HSPWM, the channel is naturally formulated as a tensor, enabling the use of tensor-OMP. Simulation results demonstrate that tensor-OMP achieves normalized mean square error (NMSE) performance comparable to SD-OMP, while offering reduced computational complexity and improved scalability.
Introduction: The Challenge of Distributed ELAA
To meet the extreme data rate demands of 6G, researchers are looking toward Extremely Large Antenna Arrays (ELAA). However, mounting a massive, monolithic array on a single platform is physically difficult. A clever workaround is to use a swarm of Unmanned Aerial Vehicles (UAVs), each carrying a smaller array, to cooperatively act as one giant, distributed ELAA.
As an antenna array grows in size, its "near-field" region (the Fresnel zone) expands dramatically. For a distributed UAV swarm, the user is almost certainly in the near-field. In this zone, the standard "Plane Wave Model" (PWM) used in traditional cellular math falls apart because it assumes signals arrive in parallel lines. While the "Spherical Wave Model" (SWM) is physically accurate, it is computationally intensive for real-time channel estimation.
Distributed ELAA System Model
The system considers a swarm of UAVs forming a distributed uniform planar array (UPA). Each UAV is equipped with its own local antenna array, and together they serve a ground user.
The geometry of this setup, as shown in Figure 1, necessitates a model that accounts for the varying distances and angles between the user and each individual UAV in the swarm.
The Hybrid Spherical-Plane Wave Model (HSPWM)
The authors propose a "Goldilocks" model—the Hybrid Spherical-Plane Wave Model (HSPWM), or cross-field model. Since each individual UAV is relatively small, the signal across its local antennas can be treated as a plane wave (PWM). However, because the UAVs are far apart, the signal across the entire swarm must be treated as a spherical wave (SWM).
A major insight from the SNR analysis is that near-field SNR does not grow infinitely as more antennas are added (as PWM falsely suggests). Instead, it is limited by the geometric aperture of the array. The HSPWM captures this physical reality perfectly, matching the accuracy of the SWM while remaining mathematically tractable.
Tensor-OMP Algorithm for Channel Estimation
Channel estimation in these systems is a sparse recovery problem. Standard SD-OMP (Spherical-Domain Orthogonal Matching Pursuit) searches a massive 3D grid of azimuth, elevation, and range, which is slow and memory-intensive.
The proposed Tensor-OMP exploits the fact that the HSPWM channel can be represented as a multi-dimensional tensor (a 3D data cube).
As illustrated in Figure 4, instead of building one giant dictionary, Tensor-OMP performs smaller operations along each dimension of the tensor. This exploits the Kronecker structure of the near-field channel, avoiding the "curse of dimensionality" and significantly slashing computational costs. To further improve accuracy, the Nelder-Mead algorithm is used for off-grid refinement, ensuring precise estimation even when the user is not perfectly aligned with the predefined grid.
Key Results and Performance
Simulation results for a 4x4 UAV swarm (2,304 total antennas) at 10 GHz demonstrate the following:
- Modeling Accuracy: The HSPWM SNR curves are nearly identical to the "perfect" SWM curves, validating the hybrid approach.
- Estimation Performance: Tensor-OMP achieves a Normalized Mean Square Error (NMSE) comparable to the more complex SD-OMP.
- Complexity Gain: Tensor-OMP significantly reduces the computational complexity of the projection step, offering a massive speedup as the number of antennas grows.
Conclusion and Future Directions
This work provides a robust framework for near-field communication in distributed UAV swarms. While the results are promising, future work must address real-world challenges such as perfect synchronization between UAVs, the need for decentralized processing, and the impact of UAV mobility and "jitter" on channel stability.
Summary updated