一类动态多尺度系统的模型和估计算法

多传感器信息融合是信息学科的重要研究方向,本文的目的就是对分布在多个尺度上的传感器信息进行融合。本文针对一类动态多尺度系统,在国内外相关研究的基础上,提出一种新的动态多尺度系统建模方法。该模型满足标准卡尔曼滤波条件,给出了基于Haar小波的实现方法,基于此可以获得各个尺度上目标状态线性最小方差意义下的最优估计值。仿真结果令人满意,最细尺度上用本文方法进行估计优于直接进行卡尔曼滤波的效果。

Model and Estimation Algorithm of a Class of Dynamic Multiscale System

Multisensor fusion is an important research area in information subject, and the purpose of this paper is to fuse the multi-sensor information that is distributed at multiscale. On the basis of the related research, a new dynamic multiscale system (DMS) modeling method of a class of dynamic multiscale system is proposed in this paper. This model satisfies standard Kalman filter conditions. The realization method based on Haar wavelet is proposed. The linear minimum mean-square error estimation at each scale can be obtained by our method. Simulation results are given to obtain insight into the efficiencies offered by our model and algorithm. At the finest scale, the estimation by our method is better than that is gained by performing Kalman filter directly.