Nonlinear filtering for terrain-referenced underwater navigation with an acoustic altimeter

Terrain-referenced navigation (TRN) uses geometric information of terrain elevations to correct drift errors due to dead-reckoning or inertial navigation. This approach can be very useful for underwater navigation, since global positioning system (GPS) signals are not available under the surface of water. However, TRN requires a geometric description of an undulating terrain surface as a mathematical function or a look-up table, which leads to a nonlinear estimation problem. Thus, the navigation performance depends on the choice of the filter algorithm. In this study, the use of an Rao-Blackwellized particle filter is considered for underwater TRN using a single-beam acoustic altimeter, and its performance is compared with two commonly used Gaussian Kalman filters (an extended Kalman filter and an unscented Kalman filter) through navigation simulations with actual bathymetry data.