Sigma Point Kalman Filter for Underwater Terrain-Based Navigation

Abstract

Precise underwater navigation is crucial in a number of marine applications. Navigation of most autonomous underwater vehicles (AUVs) is based on inertial navigation. Such navigation systems drift off with time and external fixes are needed. This paper concentrates on one such method, namely terrain based navigation, where position fixes are found by comparing measurements with a prior map. Nonlinear Bayesian methods like point mass and particle filters are often used for this problem. Such methods are often computationally demanding. The sigma point Kalman filter (or unscented Kalman filter) is a nonlinear filter that does not resort to local linearizations, estimating the probability densities using a few deterministically chosen sigma points. The sigma point Kalman filter is not as computationally demanding as the aforementioned methods, and using real AUV data, the accuracies obtained are comparable to those of the point mass and particle filters.

Development of environmental niche models for use in underwater vehicle navigation

“…The primary limitation of the technique presented in [38] was the lack of an accurate terrain map, which does not invalidate the methodology used. A number of other studies have utilised particle filters as part of a TBN framework for underwater vehicles [38][39][40][41]. The particle filter is suitable as a solution to the TBN problem because it is probabilistic (and therefore captures environmental uncertainty), and because it naturally incorporates the property that the longer a path is traversed, the more likely a single solution will emerge.…”

Development and Performance Validation of a Navigation System for an Underwater Vehicle

“…Due to the vehicle dynamics, there are possibilities that could result in data outages from the DVL that could lead to erroneous position computation. To overcome the system's degraded performance during such data outages, a linear Kalman Filter (KF) is developed and incorporated in the DVL velocity measurements (Leader, 1994; Anonsen and Hallingstad, 2007; Welch and Bishop, 1995; Simon, 2006; Mandt et al, 2001). The prediction methodology of the KF is shown below.…”

“…To overcome the system's degraded performance during such data outages, a linear Kalman Filter (KF) is developed and incorporated in the DVL velocity measurements (Leader, 1994;Anonsen and Hallingstad, 2007;Welch and Bishop, 1995;Simon, 2006;Mandt et al, 2001). The prediction methodology of the KF is shown below.…”

Robust particle filter based on Huber function for underwater terrain‐aided navigation

“…One is the batch correlation methods like TERCOM [12], iterated closest contour point (ICCP) [13], and maximum likelihood estimation [14]. The other is the recursive Bayesian methods such as Kalman filter (KF) [15], point mass filter (PMF) [16][17][18], and particle filter (PF) [19][20][21]. Compared with the batch correlation methods, the recursive Bayesian methods are more sophisticated since they incorporate motion uncertainty between adjacent measurements.…”