基于现代时间序列分析方法的通用信息融合白噪声估值器

【作者】 王世刚；

【导师】 邓自立；

【作者基本信息】 黑龙江大学 ， 控制理论与控制工程， 2008， 硕士

【摘要】 多传感器信息融合是多维信息综合处理的一项新技术,广泛应用于信息获取与处理领域,己成为当前信息领域的一个十分活跃的研究热点。随着科学技术的发展,单一传感器检测技术已不能满足要求,多传感器融合技术应运而生。多传感器融合技术就是对同一检测对象,利用各种传感器检测的信息和不同的处理方法以获得该对象的全面检测信息,从而提高检测精度和可靠性。输入白噪声估计也叫白噪声反卷积,在石油地震勘探领域中有重要的应用背景。对带不同局部模型多传感器线性离散定常随机系统,根据按标量加权最优融合规则,应用现代时间序列分析方法,基于自回归滑动平均(ARMA)新息模型和白噪声估计理论,提出一种稳态最优信息融合公共输入白噪声反卷积估值器算法。并将现代时间序列分析方法与经典Kalman滤波方法相结合,提出稳态最优信息融合公共输入白噪声稳态反卷积估值器的两种等价算法。它们的精度高于每个局部估值器的精度,且可统一处理融合滤波、预报和平滑问题。为计算最优加权,提出计算局部估计误差互协方差公式。它们可应用于解决带有色观测噪声多传感器系统白噪声反卷积融合估计问题。大量输入白噪声的Monte-Carlo的仿真例子说明了它们的有效性和正确性。更多还原

【Abstract】 Multisensor information fusion is a new technology to process multidimensional information synthetically .It has been applied in the field of information acquisition and processing widely and has become an active research focus in current information domain. With the development of science and technology, a single sensor detection techonology can not satisfy the requirements of detection, so multi-sensor fusion technology came into being. Multi-sensor fusion technology is a technology of detecting a same object, by using all kinds of detection information and different detection approach to object’s comprehensive information ,so, improves the accuracy and reliability.Input white noise is also called white noise deconvolution . it has an important applied background in oil seismic exploration.For linear discrete time-invariant stochastic systems with different local models, according to optimal fusion rules weighted by scalars ,using the modern time series analysis method, based on the autoregressive moving average (ARMA) innovation model and white noise estimation theory, an algorithm of the steady state optimal information fusion common input white noise deconvolution estimators are presented. And combine the modem time series analysis method and classical Kalman filtering method, two equivalent algotithms of the steady state optimal information fusion common input white noise deconvolution estimators are presented. Their accuracy is higher than that of each local estimator, and they can handle the fused filtering, prediction, and smoothing problems in a unified framework. In order to compute the optimal weights, the formulas of computing the cross-covariances among local estimation errors of input white noise are given. Especially, they can be applied to solve the white noise deconvolution fused estimation problem for system with color observation noise . Many Monte Carlo simulation examples for input white noise show their effectiveness and correctness.更多还原