引用本文:杨旭升,赵鑫微,张文安,章东平,来晓.基于量测置信引导的渐进高斯滤波融合方法[J].控制理论与应用,2026,43(5):1123~1132.[点击复制]
YANG Xu-sheng,ZHAO Xin-wei,ZHANG Wen-an,ZHANG Dong-ping,LAI Xiao.Progressive Gaussian filtering fusion method based on measurement confidence guidance[J].Control Theory & Applications,2026,43(5):1123~1132.[点击复制]
基于量测置信引导的渐进高斯滤波融合方法
Progressive Gaussian filtering fusion method based on measurement confidence guidance
摘要点击 344  全文点击 11  投稿时间：2024-07-10  修订日期：2025-10-07
查看全文  查看/发表评论  下载PDF阅读器   HTML
DOI编号  10.7641/CTA.2025.40367
  2026,43(5):1123-1132
中文关键词  渐进高斯滤波  卡尔曼滤波器  分布式融合估计  量测不确定性
英文关键词  progressive gaussian filtering  Kalman filters  distributed fusion estimation  measurement uncertainty
基金项目  国家自然科学基金项目(62473335, W2421117), 浙江省自然科学基金白马湖实验室联合基金项目(LBMHD24F030002), 浙江省“尖兵” “领雁”研发 攻关计划项目(2024C01028), 中国博士后科学基金项目(2024M752864)资助.
作者单位E-mail
杨旭升 浙江工业大学信息工程学院 xsyang@zjut.edu.cn 
赵鑫微 浙江工业大学信息工程学院  
张文安* 浙江工业大学信息工程学院 wazhang@zjut.edu.cn 
章东平 中国计量大学信息工程学院  
来晓 中控技术股份有限公司  
中文摘要
      针对复杂量测噪声下分布式融合估计问题, 本文提出了一种基于量测置信引导的渐进高斯滤波融合方法. 首先, 采用卡方检验法对多传感器量测进行置信分类, 以满足复杂量测噪声的分类处理要求. 其次, 设计渐进更新过 程的交互控制策略来实现局部估计的间接补偿, 同时, 基于假设检验方法给出噪声协方差的保守上界, 以获得保守 性的局部估计. 此外, 利用QR分解方法导出平方根型局部滤波方法, 以保证协方差的正定性以及提高对数值计算误 差的稳定性. 最后, 设计异步状态融合估计方法来实现分层分类融合估计. 通过仿真结果验证了所提方法的有效性.
英文摘要
      Aiming at the problem of distributed fusion estimation under complex measurement noise, this paper proposes a progressive Gaussian filtering fusion method based on measurement confidence guidance. Firstly, the chi-square test method is used to classify the multi-sensor measurements to meet the classification requirements of complex measurement noise. Secondly, the interactive control strategy of the progressive update process is designed to realize the indirect compensation of the local estimation. At the same time, the conservative upper bound of the noise covariance is given based on the hypothesis test method, so as to obtain the conservative local estimation. In addition, the square root local filtering method is derived by using the QR decomposition method to ensure the positive definiteness of the covariance and improve the stability of the numerical calculation error. Finally, an asynchronous state fusion estimation method is designed to realize hierarchical classification fusion estimation. The effectiveness of the proposed method is verified by simulation results.