子波域一种新的多尺度多传感器状态估计  

A New Method of Multi-scale Multi-sensor State Estimation in Wavelet Domain

作　　者：杨友明[1] 李建勋[1] 肖艳军[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200030

出　　处：《上海交通大学学报》2007年第11期1745-1748,共4页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金资助项目(60304007);上海市科技启明星资助计划(04QMX1410)

摘　　要：研究了多尺度状态参数估计问题.利用状态信号的多尺度特性,引入小波变换,建立了一个新的多尺度多传感器滤波算法.将函数全变差和局部变差概念引入到小波变换,提出了小波缩减因子的概念.通过小波缩减因子的引入,细尺度上的高频信号也得到了有效处理.仿真结果表明,该算法充分利用信号的多尺度特性,基于尺度级的信息融合大大提高了状态估计的精度.The problem of multi-scale parameter estimation was studied. Based on the idea of multi-scale characteristic of signal and wavelet transform, a new multi-scale multi-sensor filtering algorithm was proposed. The concepts of ＂total variation＂ and ＂local variation＂ of functions are introduced to the areas of wavelet transform, and then the concept of ＂contraction coefficients＂ is raised which significantly augments the previous theory that the updating of state estimation merely focuses on the approximations of wavelet decompositions. The Monte Carlo simulation reveals that the accuracy of estimation can be improved notably under the new theory.

关 键 词：小波变换 全变差 多尺度 状态估计 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]