Marginal Reduced High-degree CKF and its Application in Nonlinear Rapid Transfer Alignment

Abstract

The error model is not perfect when the nonlinear transfer alignment is applied to the inertial navigation system of shipborne aircraft. In this paper, a new model incorporating acceleration error of lever-arm is proposed. Considering the influence of flexure deformation and dynamic lever-arm, it is suitable for the case of large azimuth misalignment The state filtering estimation selects the High-degree Cubature Kalman filter (HCKF). A reduced HCKF algorithm with marginalized-sampling-based cubature transformation is designed according to the state equation and measurement equation structure of the transfer alignment model, and it is proved. This algorithm simplifies the measurement update process, and applies the marginal sampling algorithm to the time update process, achieving the purpose of reducing the calculation amount of HCKF. In the simulation, the velocity plus attitude matching method is used. The experimental results show that compared with the transfer alignment model without considering the dynamic lever-arm, this model has a higher accuracy, and the alignment precision and time also meet the requirements.