基于随机加权估计的Sage自适应滤波及其在导航中的应用  

Random weighting Sage adaptive filter and its application to integrated system

作　　者：侯俊[1] 朱长青[1,2] 阎海峰[1,3] 赵岩[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072 [2]空军装备部科研订货部机载处,北京100843 [3]中航工业集团深圳南航电子工业有限公司,广东深圳518052

出　　处：《电子设计工程》2014年第20期1-5,共5页Electronic Design Engineering

基　　金：国家自然科学基金资助(61174193)

摘　　要：为了克服Kalman滤波和Sage自适应滤波的缺点,在分析基于新息向量、残差向量和状态改正数向量的自适应协方差估计存在问题的基础上,提出根据新息向量、残差向量和状态改正数对滤波精度影响的不同程度,采用随机加权法对新息向量、残差向量和状态改正数进行估计,以得到观测噪声协方差矩阵和系统动态噪声协方差矩阵。进一步,利用随机加权法对观测噪声协方差阵和系统噪声协方差阵进行估计,以提高动态导航定位的滤波解算精度。研究结果表明,基于随机加权估计的Sage自适应滤波效果明显优于基于算术平均值估计的滤波方法。In order to overcome the shortcoming of Kalman filter and Sage adaptive filter, based on the analysis of the new rate vector, the residual vector and the number of state correction vector adaptive covariance estimation problems, this paper presents a new Sage adaptive filtering method. according to the new rate vector, the residual vector and the number of state correction to the influence of filtering precision degrees, this method takes the random weighting method to estimate the new rate vector, the residual vector and the number of correct state, and gets the observation noise covariance matrix and the system dynamic noise covariance matrix. Further, for improving the filter precision of calculating dynamic navigation and positioning,the proposed filter uses the random weighting method to estimate observation noise covariance matrix and the system noise covariance matrix. The results show that based on the random weighting estimation Sage adaptive filter is obviously better than arithmetic mean filtering method.

关 键 词：导航定位 卡尔曼滤波 随机加权估计 噪声协方差阵 

分 类 号：TN96[电子电信—信号与信息处理]