Model-based 3D micro-navigation and bathymetry estimation for interferometric synthetic aperture sonar

Sub-wavelength navigation information is vital for the formation of all synthetic aperture sonar (SAS) data products. This challenging requirement can be achieved using the redundant phase centre (RPC) or displaced phase centre antenna (DPCA) micro-navigation algorithm, which uses cross-correlation of signals with inter-ping coherence to estimate time delays and hence make navigation estimates. In this paper a new approach to micro- navigation for interferometric synthetic aperture sonar is introduced. The algorithm makes 3D vehicle position estimates for each sonar ping by making use of time delays measured between all possible pairs of redundant phase centre arrays, using both interferometric arrays on each side of the vehicle. Simultaneous estimation of coarse bathymetry allows the SAS images to be projected onto ground-range. The method is based on non-linear minimization of the difference in modelled and measured time delays and surges between redundant phase centre arrays. The approach is demonstrated using data collected by the CMRE MUSCLE AUV using its 270-330 kHz SAS during the MANEX’14 experiment. SAS images have been projected onto the coarsely estimated bathymetry, and interferograms have been formed. The coarse bathymetry estimate and vehicle navigation estimate are validated by the quality of the image focussing and the near-zero phase of the interferogram. The method has the potential to improve through-the-sensor navigation aiding and to increase the accuracy of single-pass bathymetry estimation. Future development of the algorithm for repeat-pass operation has the potential to enable repeat-pass track registration in three dimensions. The method is therefore an important step towards improved coherent change detection and high resolution bathymetry estimation.