基于高斯混合模型的抗差Bayes滤波在动态导航中的应用

发布时间：2018-09-10 18:51

【摘要】：在动态环境下导航定位受到很多因素影响,且异常噪声会严重影响导航滤波结果。通过研究Bayes滤波的抗差方法,在导航观测方程中采用高斯混合模型,对混合模型不确定度参数采用层次模型实时估计。借助指示变量进行模型变换,削弱导航过程中异常噪声的影响。并采用Rao-Blcakwellized粒子采样方法,求取复杂、非标准形式的状态后验分布。最后通过卫星导航以及组合导航实验算例,分析验证了基于高斯混合模型的Bayes滤波在动态导航定位中的抗差性能。
[Abstract]:In dynamic environment, navigation location is affected by many factors, and abnormal noise will seriously affect the navigation filtering results. By studying the robust method of Bayes filtering, Gao Si mixed model is used in navigation observation equation, and real-time hierarchical model is used to estimate the uncertainty of mixed model. The influence of abnormal noise in navigation process is weakened by using indicator variable to transform the model. The complex, non-standard state posterior distribution is obtained by Rao-Blcakwellized particle sampling method. Finally, the robust performance of Bayes filter based on Gao Si hybrid model in dynamic navigation and positioning is analyzed and verified by the experimental examples of satellite navigation and integrated navigation.
【作者单位】：信息工程大学;
【基金】：国家自然科学基金项目(41274016;40974010) 地理信息工程国家重点实验室重点基金项目(SKLGIE2014-Z-2-1) 信息工程大学地理空间信息学院创优基金(XS201504)
【分类号】：P228.4

【相似文献】