Lie group based nonlinear error models for SINS/DVL integrated navigation
Jun Tang,Hong-wei Bian,Rong-ying Wang,Heng Ma
DOI: https://doi.org/10.1016/j.oceaneng.2024.116898
IF: 5
2024-02-03
Ocean Engineering
Abstract:The traditional error model for strapdown inertial navigation system (SINS)/Doppler velocity log (DVL) integrated navigation is establish for special orthogonal group Kalman filter (SOKF), which is only suitable for the case of small misalignment angle. In this article, the Lie group based nonlinear error model is investigated for SINS/DVL in inertial frame, this model is no longer limited by the size of misalignment angle like traditional SOKF, so initial alignment and integrated navigation could be combined into one stage, and the navigation solution is more accurate and faster. First, a thorough theoretical analysis is presented for state transformation extended Kalman filter (STEKF) modeling in both special orthogonal group (SO(3)) and special euclidean group (SE(3)). Then, the nonlinear error models (both left and right invariant error models) in inertial frame are derived in extended special euclidean group (SE 2 (3)). On this basis, SINS/DVL is verified as a type of left invariant observation system which is more applicable to the left invariant error model. Finally, both simulation and field test for SINS/DVL are conducted to evaluate the actual performance of the investigated methods. The results show that the proposed left invariant error model in inertial frame performs better for SINS/DVL than right invariant error model, and it also has more excellent performance than other existing integrated navigation methods, such as SOKF and STEKF. Especially compared with the current advanced invariant error model defined in earth frame, the model we investigated is more concise, efficient and accurate.
engineering, civil, ocean, marine,oceanography