A fast alignment of marine strapdown inertial navigation system based on adaptive unscented Kalman Filter

Abstract

This study has presented an efficient adaptive unscented Kalman filter (AUKF) with the new measurement model for the strapdown inertial navigation system (SINS) to improve the initial alignment under the marine mooring conditions. Conventional methods of the accurate alignment in the ship’s SINS usually fail to succeed within an acceptable period of time due to the components of external perturbations caused by the movement of sea waves and wind waves. To speed up convergence, AUKF takes into account the impact of the dynamic acceleration on the filter and its gain adaptively tuned by considering the dynamic scale sensed by accelerometers. This approach considerably improved the corrections of the current residual error on the SINS and decreased the influence due to the external perturbations caused by the ship’s movement. Initial alignment algorithm based on AUKF is designed for large misalignment angles and verified by experimental data. The experimental test results show that the proposed algorithm enhanced the convergence speed of SINS initial alignment compared with some state-of-the-art existing approaches.

Experimental Research on Shipborne SINS Rapid Mooring Alignment with Variance-Constraint Kalman Filter and GNSS Position Updates

“…The second type aims to increase the tolerance of the initial attitude error in the Kalman Filter by establishing a large misalignment angle error model. Nonlinear filter methods, such as the extended Kalman filter (EKF), the unscented Kalman filter (UKF), or the cubature Kalman filter (CKF), are employed for fine alignment [ 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 ]. Although these methods do not require a small initial misalignment angle assumption, the mathematical models involved are relatively complex, and convergence may take longer in the presence of a large initial azimuth error [ 19 ].…”

A GNSS Aided Initial Alignment Method for MEMS-IMU Based on Backtracking Algorithm and Backward Filtering

“…These algorithms are used to calculate a coarse attitude, and the fine initial attitude will be estimated after modeling the error model of SINS, estimating attitude errors of coarse aligning result, and analysing the state observablity degree. The error model is mainly divided into little-misalignment inertial model [5] and large-misalignment nonlinear model [6], [7], [8]. UKF (Unsented Kalman Filter) is usually used to estimate attitude error of nonlinear model, because large-misalignment error model of SINS is strong-nonlinearity.…”