A Fault-Tolerant Polar Grid SINS/DVL/USBL Integrated Navigation Algorithm Based on the Centralized Filter and Relative Position Measurement

Abstract

Navigation is a precondition for ocean space vehicles to work safely in polar regions. The traditional polar algorithms employ the grid strapdown inertial navigation system (SINS) as the backbone and Doppler velocity log (DVL) output velocity as measurements to constitute the integrated navigation system, of which, however, the position errors still accumulate with time. The ultra-short baseline (USBL) position system can provide position information that can be used to improve the performance of the SINS/DVL integrated system. Therefore, a grid SINS/DVL/USBL integrated algorithm for polar navigation is proposed in this paper. In order to extend the availability of the USBL and improve integration accuracy in polar regions, the USBL observation model is established based on the relative position measurement firstly. Then, a grid SINS/DVL/USBL integrated algorithm is proposed to fuse the information of these sensors with a modified Kalman filter (MKF) dealing with the sparse USBL output. Finally, a vector fault detection method, which takes the measurements as detection objects instead of the filter, is designed to locate the measurement fault and can be employed by the centralized filter to improve the fault-tolerant. Simulation and experiment results show that the proposed grid SINS/DVL/USBL integrated navigation system can further restrain SINS errors especially the position errors effectively. Meanwhile, the vector fault detection method can detect and isolate the fault measurements of centralized filter immediately and accurately. Therefore, the proposed fault-tolerant grid SINS/DVL/USBL integrated navigation algorithm can improve the reliability and accuracy of polar navigation for ocean space application.

An Ultra-Short Baseline Underwater Positioning System with Kalman Filtering

“…They proposed various types of array element types [ 1 , 2 , 15 , 16 , 17 , 18 ]. Many positioning methods also have been presented to improve the positioning accuracy [ 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 19 , 20 , 21 , 22 , 23 , 24 ].…”

“…This method, with the same positioning accuracy as the orthogonal array composed of eight elements, can reduce the number of primitive elements and increase the utilization of primitive elements. From the view of signal noise reduction, most systems took advantage of the adaptive filtering algorithm or adaptive residual optimization algorithm to filter signals to obtain the signal phase difference [ 4 , 10 ], but the result was not promising. Reference [ 21 ] proposed a novel non-equidistant, quaternary array.…”

“…Wang et al proposed a robust Student’s t -based Kalman filter for the strap-down inertial navigation system (SINS) and the ultra-short baseline (USBL) integration system [ 3 ], which was utilized to suppress the measurement uncertainty induced by the acoustic outliers. Reference [ 4 ] established a USBL observation model and a grid SINS/Doppler Velocity Log (DVL)/USBL-integrated navigation system to restrain the SINS errors. When the marine environment changed in the shallow sea area, the propagation speed model needed to be corrected.…”

“…In this paper, we adopted three reference methods. They are the adaptive filtering [ 10 ], adaptive residuals filtering [ 4 ], and new four-element methods [ 19 ]. In the adaptive filtering method [ 10 ], a least mean-square (LMS) algorithm is utilized to update the adaptive parameters to filter noise.…”

“…Due to the inaccurate estimation of the gradient value, the adaptive filtering is with noise. To improve the position accuracy, the adaptive residuals filtering algorithm [ 4 ] adopts adaptive residuals correction strategy, and the convergence rate is enhanced. The new four-element method adopts the new kind of array.…”