具有乘性噪声和丢包传感器网络系统的加权观测融合估计.PDF

中国青岛 第七届中国信息融合大会 Qingdao China 2015年10月 The Seventh Chinese Information Fusion Conferenc 11 具有乘性噪声和丢包传感器网络系统的加权观测 融合估计 丁健，逄崇雁，孙书利 黑龙江大学，哈尔滨 150080 摘 要：研究了具有丢包和乘性噪声的复杂多传感器网络系统的加权观测融合估计问题。不同传感器的观测丢失采用一组具 有已知概率的 Bernoulli 分布的随机变量描述。通过将乘性噪声转化为加性噪声，应用矩阵满秩分解和加权最小二乘原理， 会 给出了加权观测融合估值器。该加权观测融合估值器具有与集中式融合估值器相同的精度，即具有全局最优性。分析了算法 的稳态特性，给出了稳态存在的一个充分条件。仿真研究验证了算法的有效性。 关键词：多传感器; 观测丢失; 乘性噪声; 加权观测融合; 满秩分解 大 中图分类号：TP273 文献标识码：A 合 Weighted Measurement Fusion Estimation for Sensor Networked Systems with Multiplicative Noises and Droppings 融 Ding Jian, Pang Chongyan, Sun Shuli Heilongjiang University, Harbin 150080 息 Abstract: This paper is concerned with the weighted measurement fusion estimation problem for a complicated multi -sensor networked system with the multiplicative noises and droppings. The data dropping phenomena of different sensors are described by a group of Bernoulli distributed random variables with the known probability distributions. By transforming the multiplicative noises 信 into the additive noises, using full-rank decomposition of matrix and weighted least-square theory, the weighted measurement fusion estimators are presented. The proposed fusion estimators have the same accuracy with the centralized fusion estimator, i.e., they are global optimal. The steady-state property is analyzed and a sufficient c