Autoregressive modeling of pulse‐to‐pulse incoherent Doppler sonar data

Abstract

An innovative approach to modeling the range evolving spectral structure of pulse‐to‐pulse incoherent Doppler sonar data is presented. This approach is demonstrated on a unique multifrequency, parallel beam oceanic data set. The approach is motivated by the desire to determine the parameters of the oceanic velocity field. Under a set of assumptions the mean frequency of the Doppler spectrum can be used to infer the velocity field of the oceanic water mass on a range evolving basis. If the Doppler spectrum is symmetric and unimodal, this mean freqency may well characterize the Doppler shift from which the relative radial velocity between the sonar platform and a dense “cloud” of scatterers may be derived. However, the Doppler spectrum may be multipeaked and/or asymmetric due to the nonhomogeneous nature of the reverberation process, instrumental effects, or beam pattern structure. In this case the mean frequency may have little relation to the Doppler shift associated with the true scaterer‐platform veloci...