应用RB无迹卡尔曼滤波组合导航提高GPS重获信号后的导航精度

针对微机电-船舶惯性导航/全球定位(MEMS-SINS/GPS)组合导航系统在GPS信号中断时造成的强非线性误差及重获信号后精度变差的问题,设计了基于Rao-Blackwellised无迹卡尔曼滤波(RB-UKF)的组合导航算法。首先,基于捷联平台欧拉失准角定义了姿态误差,建立了捷联惯导系统的非线性误差传播方程。然后,针对组合导航的状态方程为非线性而量测方程呈线性的特点,设计了RB-UKF算法,在保证精度的同时降低了计算量。最后,设计了滤波算法总体结构,分别给出了GPS信号正常时和中断时组合导航滤波计算的流程。将提出的算法用于跑车实验,结果表明:在GPS失锁20s和40s再重获信号之后,使用RB-UKF算法的组合导航系统位置精度分别优于6m和7.5m,比扩展卡尔曼滤波(EKF)算法精度提高了1.5倍以上,误差收敛速度提高了1.88~16.5倍,计算量比UKF量测更新的计算量减小了41.7%。实验显示该方法显著提升了组合导航系统GPS信号中断再恢复后的滤波精度,且易于工程实现。

Precision improvement methodology for INS/GPS after GPS outage using RB-UKF

A improved Rao-Blackwellised Unscented Kalman Filter(RB-UKF)algorithm fused by MEMS-SINS/GPS(Micro-Electrical-Mechanical System/Ship's Inertial Navigation/Global Positioning System) was established when the system has become strongly nonlinear after GPS outage. Firstly, the attitude misalignment was described based on the Euler platform error angle and the nonlinear error model for the SINS was set up. Then, as the state model was nonlinear while the observation model was linear, the RB-UKF algorithm was designed to ensure the accuracy of the UKF and to reduce its amount of computation. Finally, the algorithm structure was designed and its computing procedures when the GPS was in normal or in the period of outage were given. The algorithm was used in an in-car-experiment, and the results indicate that the SINS and GPS are fused by using the method proposed in this paper, the position error is less than 6 m or 7.5 m after 20 s and 40 s GPS outage respectively, which is 1.5 times better than that of Expand Kalman Filter(EKF) algorithm. Moreover, the error convergence time by RB-UKF is 1.88 to 16.5 times faster than that of the EKF,and the amount of computation of measurement update by RB-UKF algorithm is 41.7% smaller than that of the UKF. It concludes that the proposed method can effectively reduce the navigation errors after GPS outages for MEMS-SINS/GPS systems and has advantages in project implementation.