联合H_∞滤波算法及其在组合导航系统中的应用研究

Federated H_∞ Filtering Algorithm and Its Application in Integrated Navigation System

针对联合卡尔曼滤波器对系统模型和系统噪声先验知识依赖性过强的局限性 ,基于 H∞ 滤波对系统模型和噪声等不确定性因素具有极强的鲁棒性能的特点 ,介绍了一种 H∞ 滤波的分散化技术 ,并在此基础上提出了联合H∞ 滤波算法 ,给出了信息分配方式 ,并对其性能作了详细的分析 ,指出该算法能有效解决局部滤波器测量更新之间的相关性问题。对 MIMU/ GPS/ HMR组合导航系统进行了计算机仿真 ,仿真结果表明 ,该联合 H∞ 滤波算法虽然不是最优的 ,但是其状态估计精度仍然要优于分散 H∞

Because the model uncertainties and statistics characteristics of disturbance variations have much effects on the federated Kalman filter based on the robustness of H ∞ filter decentralized H ∞ filtering and a federated H ∞ filter are proposed. The information allocation method is also given and its capability is analyzed in detail. Analytical results indicate that the method effectively solves the correlation among the local filters. As an example of application, the filter is applied to the MIMU/GPS/HMR integrated navigation system. Simulational results prove that the federated H ∞ filter estimates the state much accurate than the decentralized H ∞ filter.