Array Shape Tracking Using Active Sonar Reverberation

This paper concerns the problem of array shape tracking from clutter for a towed active sonar array. Although heading sensors can provide towed-array shape information, they are sometimes not sufficiently available or accurate to facilitate beamforming with low sidelobe levels necessary for detection of low-Doppler targets, especially in the presence of severe bottom reverberation. We propose an array shape tracking algorithm using multipath clutter returns from a sonar ping as a distributed source of opportunity in addition to observations from only two heading sensors. More specifically a maximum likelihood (ML) array shape calibration technique is devised that exploits the relationship between the spatial and Doppler frequencies of the clutter by fitting the reverberation snapshots from a single sub-CPI to the low rank spacetime clutter subspace. The ML estimates are then tracked using a Kalman filter, which incorporates the water-pulley model for array shape dynamics.