Notice of Violation of IEEE Publication Principles: Robust Adaptive Beamforming Using Support Vector Machines

Xiaoying Ren, Yingmin Wang, Tuo Guo, Qi WangIEEE Access 2020 · DOI · OpenAlex

Editorial Note: This technical summary describes the methods presented in the paper "Robust Adaptive Beamforming Using Support Vector Machines" (IEEE Access, 2020). Readers should be aware that this paper was issued a Notice of Violation of IEEE Publication Principles. An expert committee determined that the content was copied without attribution or permission from a 2007 paper by César C. Gaudes et al., titled "Robust Array Beamforming With Sidelobe Control Using Support Vector Machines."

Abstract

This paper proposes a robust adaptive beamforming method designed to mitigate the rapid performance degradation of the Minimum Variance Distortionless Response (MVDR) beamformer when steering vector mismatches occur. By reformulating beamforming as a Support Vector Regression (SVR) problem, the authors utilize Vapnik’s $\epsilon$-insensitive quadratic loss function to enhance robustness. The method assumes the array responds only within a predefined angle error range, treating responses outside this range as abnormal. This sparsity-driven approach prevents signal self-cancellation and ensures effective interference suppression. The optimization problem is solved using an Iterative Reweighted Least Squares (IRWLS) procedure, validated through computer simulations.

Introduction: The Challenge of Steering Vector Mismatch

In array signal processing, adaptive beamformers like the MVDR are designed to maximize the Signal-to-Interference-plus-Noise Ratio (SINR) by placing deep nulls in the direction of interference while maintaining a distortionless response toward the desired signal. However, these systems are highly sensitive to "steering vector mismatch"—errors arising from imprecise sensor placement or inaccurate Direction of Arrival (DOA) estimation.

When a mismatch occurs, the MVDR beamformer often misidentifies the desired signal as an interferer, leading to signal self-cancellation. This results in a catastrophic drop in performance, especially at high signal-to-noise ratios. To combat this, the authors propose a framework based on Support Vector Machines (SVM).

Methodology: SVR-Based Beamforming

The "Tube" Approach

The core of the proposed method is the application of Support Vector Regression (SVR). The beamformer is designed to maintain a specific response within a small angular "tube" around the estimated DOA.

Illustration of the $\epsilon$-insensitive tube in Support Vector Regression.

As shown in the figure above, the $\epsilon$-insensitive zone (the "tube") allows for a range of values where errors are not penalized. In the context of beamforming, this "dead zone" ignores small angular deviations within the estimated error range ($\alpha$), preventing the algorithm from over-reacting to minor mismatches. Data points falling outside this tube are penalized, which helps the beamformer distinguish between the desired signal (within the tube) and interference (outside the tube).

Quadratic Loss Function

While standard SVR often uses a linear $\epsilon$-insensitive loss function, this paper advocates for a quadratic loss function to improve the suppression of strong interference.

Comparison between linear ($m=1$) and quadratic ($m=2$) loss functions in the SVR framework.

The quadratic penalty ($m=2$) is significantly more sensitive to large errors than the linear penalty ($m=1$). In beamforming, large errors correspond to strong interference signals. By using a quadratic loss, the algorithm places a much higher priority on suppressing these "outliers," resulting in deeper nulls in the interference directions and a more robust main lobe.

Optimization via IRWLS

To solve the resulting quadratic convex optimization problem, the authors employ the Iterative Reweighted Least Squares (IRWLS) procedure. This choice is driven by computational efficiency. Unlike general-purpose quadratic programming solvers, IRWLS converges quickly to the SVM solution. By solving the problem in its "primal" form, the computational complexity scales primarily with the number of sensors ($M$) rather than the number of angular samples ($N$), making it suitable for real-time applications where $M \ll N$.

Simulation and Results

The effectiveness of the SVM-based beamformer was tested using a Uniform Linear Array (ULA) with 10 sensors. Simulations involved a desired signal at 0° and interference at 50°.

  • Robustness to Mismatch: In scenarios with a 2° steering vector mismatch, traditional MVDR performance collapsed to nearly -30 dB SINR as signal power increased. The proposed quadratic SVM beamformer (BF_quad) maintained a stable SINR between 20-25 dB.
  • Interference Suppression: The BF_quad method successfully placed deep nulls (approximately -60 dB) at the interference angle, even under mismatch conditions.
  • Convergence Speed: The algorithm demonstrated high stability with a low number of data snapshots, outperforming traditional Second-Order Cone Programming (SOCP) methods in dynamic environments.

Conclusion and Limitations

The SVM-based approach offers a powerful alternative to traditional robust beamforming by leveraging sparsity and insensitive loss functions. However, the authors note some trade-offs:

  1. Sidelobe Levels: The increased robustness can lead to higher sidelobe levels compared to non-robust methods.
  2. Parameter Sensitivity: The performance is contingent on an accurate estimation of the angular error range ($\alpha$).

Despite these limitations, the integration of SVR principles into array signal processing provides a mathematically rigorous way to handle the practical uncertainties of real-world signal environments.

Summary updated

Abstract

Notice of Violation of IEEE Publication Principles<br><br> “Robust Adaptive Beamforming Using Support Vector Machines”<br><br> by Xiaoying Ren, Yingmin Wang, Tuo Guo, Qi Wang<br> in IEEE Access, Volume 8, July 2020<br> After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE’s Publication Principles.<br><br> This paper contains content copied from the paper cited below. The original text was copied without attribution (including appropriate references to the original author(s) and/or paper title) and without permission.<br><br> “Robust Array Beamforming With Sidelobe Control Using Support Vector Machines”<br> by César C. Gaudes, Ignacio Santamaría, Javier Vía, Enrique Masgrau Gómez, Talesa Sesé Paules<br> in the IEEE Transactions on Signal Processing, Vol. 55, No. 2, February 2007<br><br> <br/> A robust adaptive beamformer based on the support vector machines is proposed to solve the rapid degradation of the performance of minimum variance distortionless response beamformer when the steering vector is mismatched. The proposed beamformer assumes that the angle error range in the steering vector can be estimated; the array only responds within this range, and the array response outside the angle error range is considered as the abnormal response. The proposed method is expressed in the form of support vector regression, and Vapnik’s <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\varepsilon $ </tex-math></inline-formula> -insensitive quadratic loss function which can improve the robustness of beamformer is used as a regular term to punish the abnormal response. The assumption that the array only responds within a certain angle range reflects the sparsity of support vector regression and improves the ability of the algorithm against steering vector mismatch because the energy conservation theorem holds that the response is large. The regular term ensures that a deep null exists in the direction of interference, the interference can be effectively suppressed. The iterative reweighted least square procedure can converge to the real solution of the SVM regression problem and is easy to calculate; thus, it is used to solve the algorithm proposed in this study. The effectiveness of the algorithm is verified through computer simulations.
BibTeX
@inproceedings{ren2020notice,
  author = {Xiaoying Ren and Yingmin Wang and Tuo Guo and Qi Wang},
  title = {{Notice of Violation of IEEE Publication Principles: Robust Adaptive Beamforming Using Support Vector Machines}},
  booktitle = {IEEE Access},
  year = {2020},
  doi = {10.1109/access.2020.3009993},
}

Loading related content...

Figure