多传感器系统模型参数和噪声统计的一种信息融合辨识方法  

An Information Fusion Identification Method for Model Parameters and Noise Statistics of Multisensor Systems

作　　者：高媛[1] 徐慧勤[1] 邓自立[1] 孟华[1] 王欣 毛琳[1] 

机构地区：[1]黑龙江大学自动化系,哈尔滨150080

出　　处：《科学技术与工程》2009年第17期4896-4900,共5页Science Technology and Engineering

基　　金：国家自然科学基金(60874063);黑龙江省教育厅科学技术项目(11521124);黑龙江省电子工程重点实验室项目(DZZD2006-16)资助

摘　　要：对于带有未知模型参数和未知相关噪声统计的多传感器随机系统,基于ARMA新息模型,利用相关方法,用平均局部的模型参数和噪声统计估值器的方法,提出了模型参数和噪声统计信息的在线信息融合估计器,它们可以被解释为最小二乘融合估计,并证明了相应的辨识器具有强一致性,即以概率1收敛于相应的真实值。一个2传感器系统的仿真例子说明其有效性。For the multisensor stochastic systems with unknown model parameters and unknown noise statistics, based on ARMA innovation model, by correlation method, using the method which makes the average of the local estimators of the model parameters and noise statistics, the online information fusion estimators of model parameters and noise statistics are presented. Which can be viewed as the least squares fused estimation, and the strong con- sistence of the corresponding identifiers is proved. That is, those identifiers converges to corresponding true values with probability one. A simulation example for a 2-sensor system shows its effectiveness.

关 键 词：信息融合估计 未知模型参数 相关方法 强一致性 辨识器 

分 类 号：O211.64[理学—概率论与数理统计]