H_∞滤波在GPS/INS组合导航系统中的应用研究

Study on Usage of H_∞ Filter in GPS/INS Integrated Navigation System

提出了 GPS/ INS位置、速度、姿态组合方法 ,并把线性时变离散系统的 H∞ 滤波应用于组合系统。由于 GPS存在多路径误差及 SA误差等原因 ,难以确定准确的噪声统计模型。H∞ 滤波对噪声的不确定具有鲁棒性 ,所以用于组合系统能取得高于 Kalm an滤波的效果。文中对基于Motion Pak惯性组件和三个 Jupiter GPS接收机组成的组合系统进行了实验研究。实验结果表明 ,H∞ 滤波取得了较好的效果 。

The Kalman filter is widely used in integrated navigation system. Based on the known noise model and plant dynamics, the Kalman filter is optimal state estimator. For most real systems, because of the influence on environment etc, the statistic model of noise doesn′t match the reality at all. H ∞ filter is robust to noise model uncertainties. GPS with more antennas can provide not only position and velocity but also heading and attitude. As influence of the multipath and SA, The errors of all the information provided by GPS are not white noise and are difficult to make precise statistic model. This paper uses the H ∞ filter of finite horizon linear time varying discrete time system to integrate GPS/INS. The experiment in integrated system is based on all the navigation information: position velocity and attitude using MotionPak IMU and three Jupiter GPS receivers. The H ∞ filter obtains better navigation precision than Kalman filter.