一种对IMM真实误差协方差的估计方法  

Approach to Estimate True Error Covariance of IMM

作　　者：曲彦文[1] 杨静宇[1] 

机构地区：[1]南京理工大学计算机科学与技术学院,江苏南京210094

出　　处：《系统仿真学报》2007年第23期5597-5600,共4页Journal of System Simulation

摘　　要：交互式多模型IMM被认为是对混合系统估计的一种性价比最好的算法之一。由于真实模式序列未知,由IMM得到的误差协方差矩阵通常与真实值不相等。在假设条件下,分析了一种新的对真实误差协方差的估计方法,并且通过Monte-Carlo仿真用真实误差协方差矩阵的平均值来评估IMM的性能。Interacting multiple model method IMM has been shown to be one of the most cost-effective algorithms for the estimation of hybrid system. The true mode sequence was not known, so the error covariance matrices which were calculated via IMM were not accurate. The effectiveness of a new approach to estimate the true error covariance was investigated under some assumptions, and the mean of true error covariance was utilized to estimate the performance of lMM via Monte-Carlo simulations.

关 键 词：混合系统 交互式多模型 卡尔曼滤波器 真实误差协方差 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]