2025_programme: Efficient Off-grid Deconvolution Beamforming for Near Field Sources in Arbitrary Arrays

Back


  • Day: June 17, Tuesday
      Location / Time: A. TERPSIHORI at 18:00-18:20
  • Last minutes changes: -
  • Session: 20. Underwater acoustic calibration, measurement, and standards
    Organiser(s): Stephen Robinson, Will Slater
    Chairperson(s): Stephen Robinson, Will Slater
  • Lecture: Efficient Off-grid Deconvolution Beamforming for Near Field Sources in Arbitrary Arrays
    Paper ID: 2332
    Author(s): Jianli Huang, Yu Wang, Zaixiao Gong, Jun Wang, Haibin Wang
    Presenter: Jianli Huang
    Abstract: Near-field source localization plays a crucial role in underwater acoustic detection. The spherical wavefront model is commonly used for near-field signal propagation. However, compared with far-field planar wavefront model, near-field source localization based on the spherical wavefront model requires the addition of a distance dimension in the searching grid. A uniform searching strategy, similar to the far-field case, generates a large number of grid points and significantly increases computational complexity. To address this issue, a deconvolution-based near-field beamforming method in beam domain is proposed, incorporating a non-uniform searching strategy in both angle and distance. Firstly, we derive a non-uniform searching grid for distance in the near-field region based on the array's effective aperture. In the angular dimension, only appropriate sampling is required in mainlobe region of each beam covering the entire angular range. This strategy reduces unnecessary sampling grids and lowers the size of the steering vector dictionary, thereby improving computational efficiency. Furthermore, a generalized near-field deconvolution beamforming model in beam domain is constructed for arbitrary arrays. An off-grid sparse Bayesian learning algorithm is then employed to solve the model, which enhances source localization accuracy. Simulation results demonstrate that the proposed method preserves the estimation accuracy while significantly improving computational efficiency especially for large-scale arrays. The method has substantial application potential, providing an effective solution for near-field source localization in underwater acoustic detection.
  • Corresponding author: Ms Jianli Huang
    Affiliation: State Key Laboratory of Acoustics, Institute of Acoustics, Chinese Academy of Sciences
    Country: China

Back