Out-of-band beamforming and matched field processing

Beamforming and matched field processing (MFP) are array signal-processing techniques for estimating wave propagation directions and source locations, respectively. Both techniques can be formulated as spatial filtering operations that weight and combine the array recordings. In conventional plane- and spherical-wave beamforming, the weights are determined from analytical field models for the corresponding wave type. In conventional MFP, the weights are commonly determined from a computational field model that accounts for known propagation complications (reflection, refraction, scattering, diffraction, etc.) in the acoustic environment. Conventional MFP reduces to conventional beamforming when the environment is simple enough to be described by free-space propagation. Even at high signal-to-noise ratios, both techniques have limitations set by the recorded frequencies, the spacing of the array elements, the geometrical extent of the array, and mismatch between the recorded and modeled acoustic fields. Interestingly, these limitations can be overcome through signal processing techniques that recover out-of-the-signal band information from in-the-signal-band recordings. Here, in-band and out-of-band beamforming and MFP results are illustrated and compared using propagation simulations in free-space and multipath environments, and array recordings from a laboratory water tank and ocean propagation experiments. [Sponsored by the Office of Naval Research and the National Science Foundation.]Beamforming and matched field processing (MFP) are array signal-processing techniques for estimating wave propagation directions and source locations, respectively. Both techniques can be formulated as spatial filtering operations that weight and combine the array recordings. In conventional plane- and spherical-wave beamforming, the weights are determined from analytical field models for the corresponding wave type. In conventional MFP, the weights are commonly determined from a computational field model that accounts for known propagation complications (reflection, refraction, scattering, diffraction, etc.) in the acoustic environment. Conventional MFP reduces to conventional beamforming when the environment is simple enough to be described by free-space propagation. Even at high signal-to-noise ratios, both techniques have limitations set by the recorded frequencies, the spacing of the array elements, the geometrical extent of the array, and mismatch between the recorded and modeled acoustic fields. In...