AUV Terrain-Aided Navigation using a Doppler Velocity Logger★

Abstract

The potential of terrain-aided navigation (TAN) of autonomous underwater vehicles (AUVs) has been demonstrated in the last ten years in a series of trials where high localization accuracy was obtained during long periods of time in different types of terrain. Despite these recent advances, it is recognized by the navigation community that further research is necessary to transform TAN into a mature navigation methodology. The present paper describes the work developed towards the development of terrain navigation methods for small AUVs, relying on standard navigation sensors and dispensing with the need of dedicated sensors for terrain data acquisition. The research described addresses the problem of implementation of terrain navigation in underwater scenarios characterized by smooth sea-bottom topography and very shallow water, where the terrain information available for navigation is scarce. The TAN algorithms and the data fusion methods whose tests are documented in the paper build upon prior theoretical work published by the authors. The experimental results reported were obtained recently through trials in the water performed with an autonomous surface vehicle acting as a proxy to an AUV.

An INS/Geomagnetic Integrated Navigation Algorithm Based on Matching Strategy and Hierarchical Filtering

“…Based on the filtering results of the geomagnetic and inertial subsystem, the main filter is established according to Formula (11) and (12). Other simulation parameters are the same as the geomagnetic filter subsystem.…”

“…Other simulation parameters are the same as the geomagnetic filter subsystem. Figure 10 gives the filtering results of the integrated navigation system: Based on the filtering results of the geomagnetic and inertial subsystem, the main filter is established according to formula (11) and (12). Other simulation parameters are the same as the geomagnetic filter subsystem.…”

“…As for the tight navigation algorithm, the measured magnetic values are directly used as observations of the integrated navigation filter [6]. As the measurement equation of tight navigation system is always nonlinear [7,8], the Extended Kalman Filter (EKF) [9], Unscented Kalman Filter (UKF) [10] and Particle Filter (PF) [11][12][13] are most commonly used. A linear approximation of the geomagnetic map is needed in EKF and UKF, but not for PF.…”

Terrain‐aided navigation for long‐endurance and deep‐rated autonomous underwater vehicles

“…Therefore, the terrain measurements are characterised by nonlinear functions of the vehicle position. If conditions for successful use of a KF do not hold (even approximately), as for the application considered in this paper (instead of an expensive multibeam echo‐sounder, a low information sonar is considered) and in Donovan () and Teixeira, Quintas, & Pascoal (), nonparametric and nonlinear Bayesian estimators need to be used. This has led to the application of PFs (Arulampalam, Maskell, Gordon, & Clapp, ).…”

“…As an alternative to using an energy‐expensive dense‐ranging multibeam sonar and high‐grade INS, this paper reports the use of simple dead‐reckoning sensors and altitude measurements from an acoustic doppler current profiler (ADCP) when in bottom‐tracking (BT) mode. By using an ADCP to measure the distance to the bottom while estimating AUV velocity (Donovan, ; Teixeira et al, ), significant power can be conserved as there is no need for an additional active sensor dedicated to measure vehicle altitude.…”