自校正分布式观测融合Kalman估值器  

Self-tuning measurement fusion distributed Kalman estimators

作　　者：李云[1] 郝钢[2] 张玉茹[1] 

机构地区：[1]哈尔滨商业大学计算机与信息工程学院,哈尔滨150028 [2]黑龙江大学电子工程学院,哈尔滨150080

出　　处：《哈尔滨商业大学学报（自然科学版）》2011年第1期94-99,共6页Journal of Harbin University of Commerce：Natural Sciences Edition

摘　　要：对于带未知噪声统计的多传感器系统,应用加权最小二乘(WLS)法得到了一个加权融合观测方程,且它与状态方程构成一个等价的观测融合系统.应用现代时间序列分析方法,基于观测融合系统的滑动平均(MA)新息模型参数的在线辨识,可在线估计未知噪声方差,进而提出了一种加权观测融合自校正Kalman估值器,可统一处理自校正融合滤波、预报和平滑问题,并证明了它的收敛性,即若MA新息模型参数估计是一致的,则它与相应的最优加权观测融合Kalman估值器的误差收敛到零,因而具有渐近全局最优性.一个带3传感器跟踪系统的仿真例子说明了其有效性.For the multisensor systems with unknown noise statistics,by the weighted least squares（WLS） method,a weighted fused measurement equation was obtained,and it together with the state equation constitute a equivalent measurement fusion system.Using the modern time series analysis method,based on on-line identification of the moving average（MA） innovation model parameters for the measurement fusion system,the on-line estimators of noise variances could be obtained,and a self-tuning weighted measurement fusion Kalman estimator is presented,which can handle the self-tuning fused filtering,prediction,and smoothing problems in a unified framework.The convergence was proved,i.e.if the parameter estimation of the MA innovation model was consistent,the error between it and optimal weighted measurement Kalman estimator converges to zero,so that it had asymptotic global optimality.A simulation example for a tracking system with 3-sensor showed its effectiveness.

关 键 词：多传感器 加权观测融合 自校正Kalman估值器 辨识 噪声方差估计 现代时间序列分析方法 渐近全局最优性 

分 类 号：O211.64[理学—概率论与数理统计]