Mode filtering approaches to acoustic source depth discrimination

Mode filtering in passive sonar represents a class of linear beamforming approaches in which the steering vectors are parameterized by the depth-dependent mode functions and horizontal wavenumbers that characterize propagation in a shallow waveguide. In this work, a mode-filter beamformer is combined with different decision metrics for the purpose of classifying the depth of an acoustic source. The concept is based on the premise that deep and shallow sources robustly separate in mode space. Experimental results using SwellEx-96 vertical line array data is shown. Performance dependence on parameters such as frequency and aperture sampling is discussed.