The DOA Estimation Method for Low-Altitude Targets under the Background of Impulse Noise

Bin Lin, Guoping Hu, Hao Zhou, Guimei Zheng, Yuwei Song2022 · DOI

TL;DR

This paper introduces a novel Direction of Arrival (DOA) estimation method for low-altitude targets operating in environments with impulse noise and multipath effects. By combining spatial differencing with a modified projection subspace algorithm, the method effectively mitigates these challenges, leading to improved accuracy compared to existing techniques without requiring prior knowledge of the noise characteristics.

What Problem Does This Paper Address?

Detecting and localizing low-altitude targets, such as drones or vehicles flying close to the ground, presents a significant challenge in radar and signal processing. These environments are often characterized by complex terrain, leading to multipath propagation where signals bounce off surfaces before reaching the receiver. This creates correlated direct and reflected signals, which can degrade the performance of standard DOA estimation algorithms, leading to angle loss. Furthermore, the presence of impulse noise, common in such environments due to factors like atmospheric conditions or electronic interference, further complicates signal processing. Impulse noise is characterized by its non-Gaussian nature and "heavy tails," meaning it can produce sudden, large amplitude spikes that are difficult for traditional algorithms, often designed for Gaussian noise, to handle. Existing methods for dealing with impulse noise often require prior knowledge of the noise's statistical properties (like the characteristic exponent of alpha-stable distributions), which is not always available in real-world scenarios. This paper aims to address the combined issues of multipath effects and impulse noise for low-altitude target DOA estimation, seeking a robust and accurate solution that doesn't rely on detailed noise characterization.

Methods

The proposed method integrates two key techniques: spatial differencing and a modified projection subspace algorithm.

  1. Spatial Differencing: To combat the correlated nature of direct and reflected signals caused by multipath, the received signal from the array is divided into overlapping sub-arrays. The spatial difference algorithm then computes the difference between the covariance matrices of these sub-arrays. This process aims to decorrelate the signals and restore the covariance matrix to full rank, thereby mitigating the impact of multipath effects.
  2. Modified Projection Subspace Algorithm: After spatial differencing, impulse noise still significantly affects the covariance matrix. This algorithm first estimates the array signal vector using a least squares criterion. Based on this estimate, a modified projection matrix is constructed. This matrix is then used to modify the covariance matrix, specifically targeting and eliminating the cross-covariance terms between the signal and the noise, which are detrimental to estimation accuracy. The authors then employ a maximum likelihood criterion to find an optimal correction coefficient to further refine the covariance matrix. Finally, the adjusted covariance matrix is used with the MUSIC algorithm for DOA estimation.

A key advantage highlighted is that this method does not require prior knowledge of the impulse noise's characteristic exponent, making it more adaptable to varying environmental conditions.

Datasets

The paper does not use real-world datasets. Instead, it relies on simulations to evaluate the proposed algorithm. The simulations involve generating synthetic signals under controlled impulse noise models (Gaussian Mixture Model - GMM and Symmetric Alpha-Stable - SaS) and multipath conditions.

Key Insights and Ideas

  • Synergistic Approach: The paper demonstrates that combining spatial differencing for multipath mitigation with a modified projection subspace algorithm for impulse noise suppression is more effective than either technique alone.
  • Robustness to Unknown Noise: A significant insight is the development of a method that achieves good performance without needing to know the specific parameters of the impulse noise, which is a practical advantage over many existing algorithms.
  • Effective Impulse Noise Handling: The modified projection matrix and the subsequent covariance matrix correction effectively isolate and remove the detrimental effects of impulse noise, which is crucial for accurate DOA estimation in challenging environments.
  • Improved Low-Altitude Performance: The method is specifically tailored for low-altitude targets, where multipath effects are particularly pronounced, and shows superior performance in these scenarios.

Key Results

The simulation results indicate that the proposed algorithm achieves higher accuracy in DOA estimation for low-altitude targets compared to conventional algorithms, including ROC-SSMUSIC, FLOM-SSMUSIC, FLOC-SSMUSIC, and IN-SSMUSIC, under both GMM and SaS impulse noise models.

  • GMM Noise: The algorithm maintains high resolution and accuracy even when the probability of outliers increases (e.g., from 0.1 to 0.3).
  • SaS Noise: The method shows robustness to changes in the characteristic exponent of the SaS noise, indicating it can handle varying degrees of impulsiveness.
  • SNR Performance: The proposed method generally outperforms others across a range of Signal-to-Noise Ratios (SNRs), especially when the SNR is greater than 0 dB.
  • Snapshot Dependency: The performance of the algorithm improves with an increased number of snapshots, a common characteristic of subspace-based methods, but it still maintains a lower RMSE than other algorithms even with fewer snapshots.

Limitations and Future Work

  • Computational Complexity: While the paper claims lower computational complexity compared to some iterative methods, the complexity analysis shows it can still be significant, especially with a large number of array elements or snapshots.
  • Coherent and Incoherent Mixed Signals: The current work focuses on impulse noise. The authors suggest that future work could extend the spatial difference method to handle both coherent and incoherent mixed signals in the background of impulse noise.
  • Parameter Estimation: The paper assumes the number of sources (K) is known. Future work might explore methods that can estimate K automatically.
  • Real-World Validation: The results are based on simulations. Further validation with real-world data would be beneficial.

Summary updated

BibTeX
@misc{lin2022doa,
  author = {Bin Lin and Guoping Hu and Hao Zhou and Guimei Zheng and Yuwei Song},
  title = {{The DOA Estimation Method for Low-Altitude Targets under the Background of Impulse Noise}},
  year = {2022},
  doi = {10.3390/s22134853},
}
Figure