Non-conventional application of robust matched-field localization: small aperture array and mid-frequency signals

Conventional applications of matched-field processing (MFP) use large aperture vertical arrays and low frequency signals. It is well known that MFP’s sensitivity to environmental mismatch is proportional to frequency. Thus, real-world application of MFP to mid-frequency signals, e.g., 800 3000 Hz, is generally regarded a very difficult problem. Using small aperture vertical arrays can also compromise the performance of MFP. However, small aperture vertical arrays are more practical for real-world scenarios. In this paper, we propose the broadband L1-norm estimator for robust broadband matched-field localization of mid-frequency signals received on extremely small aperture vertical arrays. Results using a simulated Gulf of Mexico environment for broadband signals (1000 3000 Hz) received on a 3-meter vertical array demonstrate the significant performance gains in using the L1-norm estimator over the asymptotically-optimal maximum a posteriori estimator in the presence of finite environmental sampling.