| Inertial navigation system is nonlinear in nature, while the low accuracy inertial navigation system will lead to a serious non-linear. Linear error models used are designed commonly in the assumption of a small attitude error angle, and because of neglecting the high order items, when there is a big navigation error, the linear error models can not accurately describe system nonlinear characteristics. Nonlinear model for direct IMU/GPS integrated navigation issues, not only can more accurately describe the system characteristics, but also not require a lot of constraints. This paper systematically studied the nonlinear model for data processing with UKF algorithms, and successfully applied on the IMU/GPS integrated navigation system data fusion processing.The main works and contributions are summarized as follows:1. The paper introduces briefly the development of Kalman filtering, and it's current research situation both in home and abroad.2. Discussed the principle of Kalman filter and nonlinear UKF recursive algorithm. And for the system state noise and observation noise is additive, improved UKF algorithms proposed.3. According to calculation of the actually measured data, indicated that for a strong nonlinear IMU/GPS integrated navigation system data processing, selecting UKF method has more advantages than EKF (IEKF).4. Attitude error angles'impact on the navigation solution calculation and analysis of measured data is compared and analysed. The results shown that attitude error angle values can generally increased information usage efficiency of the model, and Kalman filter results calculated diverge easily without the attitude angle; but in the case of adding measurement information from the GPS attitude error angle can limit the model error much better, make the results converged faster, and improve the accuracy of the calculation.5. Designing the IMU/GPS integrated navigation system unscented Kalman filter, and checking the calculation results with the measured data.
|
PDF Full Text Request