EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Embedded Mixture-Correntropy Spatial Smoothing for Robust DOA Estimation in Shallow-Water Underwater Acoustics

Guanquan Dai, Yang Shi, Fei-Yun Wu

Journal of Marine Science and Engineering · 21.05.2026 · 14 (10) · 957

DOI: 10.3390/jmse14100957 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Marine Science and Engineering, online erschienen am 21.05.2026. Verfasst von Guanquan Dai und 2 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China, National Natural Science Foundation of China, Fujian Provincial Department of Science and Technology.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Direction-of-arrival (DOA) estimation in shallow-water underwater acoustics is challenged by coherent multipath and impulsive disturbances, which jointly cause covariance rank deficiency and outlier-driven subspace distortion. This paper proposes an embedded robust covariance-construction mechanism for coherent-plus-impulsive DOA estimation. The mechanism is implemented as mixture-correntropy-weighted simplified spatial smoothing (SS–MCC), in which snapshot reliability is enforced during subarray covariance accumulation rather than after decorrelation. A two-kernel residual-based weighting rule suppresses strongly contaminated snapshots while retaining moderately perturbed but informative snapshots. Under a controlled narrowband uniform linear array benchmark with fully coherent two-arrival multipath and Bernoulli–Gaussian impulsive noise, SS–MCC yields more stable DOA behavior than MUSIC, SS-MUSIC, and FLOM-MUSIC, especially in low-SNR, high-impulsiveness, and near-threshold regimes, although absolute strict recovery remains limited in the hardest cases. All-trial strict correct-two-peak statistics and ablation results show that the gain mainly comes from embedded covariance cleaning rather than post-processing or parameter tuning. A measured-noise-injected benchmark using NOAA–Navy SanctSound FK01 underwater recordings further confirms the same qualitative robustness trend under real noise waveforms, while remaining a semi-realistic noise-injection check rather than measured-array sea-trial validation. A simplified DOA- assisted MVDR benchmark indicates that improved covariance robustness can also support more favorable beamforming-oriented trends. The results provide controlled benchmark evidence that reliability-aware covariance construction can stabilize subspace extraction under joint coherent multipath and impulsive contamination; validation under wideband propagation, model mismatch, partial coherence, and measured array data remains future work.

Förderung

  • National Key Research and Development Program of China
  • National Natural Science Foundation of China
  • Fujian Provincial Department of Science and Technology

Zitieren

APA

Dai, G., Shi, Y., & Wu, F. (2026). Embedded Mixture-Correntropy Spatial Smoothing for Robust DOA Estimation in Shallow-Water Underwater Acoustics. Journal of Marine Science and Engineering, 14(10), 957. https://doi.org/10.3390/jmse14100957

Vancouver

Dai G, Shi Y, Wu F. Embedded Mixture-Correntropy Spatial Smoothing for Robust DOA Estimation in Shallow-Water Underwater Acoustics. Journal of Marine Science and Engineering. 2026;14(10):957. doi:10.3390/jmse14100957

BibTeX
@article{Dai2026,
  author = {Dai, Guanquan and Shi, Yang and Wu, Fei-Yun},
  title = {Embedded Mixture-Correntropy Spatial Smoothing for Robust DOA Estimation in Shallow-Water Underwater Acoustics},
  journal = {Journal of Marine Science and Engineering},
  year = {2026},
  volume = {14},
  number = {10},
  pages = {957},
  doi = {10.3390/jmse14100957}
}
Kostenlos starten →