Research on Nonlinear Comprehensive Calibration Algorithm for the Single-Axis Rotation Inertial Navigation System Based on Modified Unscented Kalman Filter

Authors:
  • JCJianhua Cheng
    Jianhua Cheng
  • MLMeiling Li
    Meiling Li
  • DGDongxue Guan
    Dongxue Guan
  • TWTongda Wang
    Tongda Wang
  • WZWei Zhang
    Wei Zhang

Abstract
Single-axis rotation inertial navigation system (SRINS) does not have the ability to achieve the modulation of all error sources of inertial navigation system (INS) by rotation. This would cause the problem that the residual error sources of SRINS will cause the navigation errors to accumulate over time, and lead to further evident nonlinear characteristics. Therefore, the traditional linear filter is unable to achieve the optimal estimation for the nonlinear error models any more. As the problem arise when the linearization of error models and linear filters are applied in nonlinear comprehensive calibration, this paper innovatively applies unscented Kalman filter (UKF) to SRINS by introducing the external position reference, and achieves the comprehensive calibration method of nonlinear filter estimation. Besides the nonlinear characteristic, the most important problem for the calibration is how to shorten the estimation time because of the poor observability of the azimuth gyroscope drift. Therefore, the modified UKF is designed to enhance the filter stability by introducing ΔQ k , meanwhile guaranteeing accuracy of states estimation. The nonlinear comprehensive calibration of SRINS could be achieved by resetting navigation information, and the working cycle of SRINS is prolonged to satisfy long term position endurance. Simulation results show that the position endurance levels are compatible with 0.8 nmile/24 h, which means that this method can realize the comprehensive calibration, and verify the effectiveness and feasibility of the calibration algorithm.