文章摘要
戴文舒,陈新华,孙长瑜,余华兵.等价权修正的Kalman抗野值滤波算法[J].,2013,32(5):409-412
等价权修正的Kalman抗野值滤波算法
Modified Kalman Filter Algorithm by Equivalent Weights With Immunity To Outliers
投稿时间:2013-07-23  修订日期:2013-08-26
中文摘要:
      水声工程实际数据中野点的存在影响标准Kalman滤波精度。借鉴抗差惯性导航定位中等价权思想,构造反映观测值误差的统计量,以其协方差矩阵作为判断滤波精度的指标,依据莱特准则设计了三种不同权重因子,充分利用正确观测,限制利用可用观测,排除有害粗差。仿真结果表明,这种等价权修正Kalman抗差算法,抗差性可媲美中位数滤波,证明当模型正确时,其在抑制零星和斑点型野值方面,效果很好。
英文摘要:
      The equivalent weighting algorithm of robust estimation for inertial navigation and positioning is applied to the standard Kalman filter when the outliers existed in the actual data. Statistic that reflects observation error is constructed with its covariance matrix acting as the indicator of the filtering accuracy. Thus the new algorithm uses the most of the correct observations, limits the median quality observations and removes harmful gross errors. Simulation results showed that the robustness is compatible with the median filtering. When the model is correct, the modified filtering algorithm is efficiently resistant to scattered and spotted outliers.
DOI:10.11684/j.issn.1000-310X.2013.05.013
中文关键词: Kalman滤波  等价权  中位数滤波  抗差  野值
英文关键词: Kalman filter  Equivalent Weights  Median filtering  robust estimation  Outliers
基金项目:
作者单位E-mail
戴文舒* 中科院声学所 dws.01@163.com 
陈新华 中科院声学所 cxh@ocean.ioa.ac.cn 
孙长瑜 中科院声学所 sunchangyu@mail.ioa.ac.cn 
余华兵 中科院声学所 yuhuabing@mail.ioa.ac.cn 
摘要点击次数: 2626
全文下载次数: 3220
查看全文   查看/发表评论  下载PDF阅读器
关闭