节点文献
基于Riccati方程的自校正信息融合滤波方法研究
【作者】 王强；
【导师】 邓自立；
【作者基本信息】 黑龙江大学 ， 控制理论与控制工程， 2009， 硕士
【摘要】 多传感器信息融合滤波重要方法之一是利用多个传感器对同一目标进行检测,从而获得其状态的局部估计,并在一定最优融合准则下,组合或加权局部估计,从而获得最优融合估计,其精度要比每一个局部估计更精确。自校正信息融合滤波是用来处理含未知模型参数和噪声统计多传感器系统的信息融合滤波问题,它是最优信息融合滤波与系统辨识两个科学的交叉,具有重要理论和应用意义。对带有未知模型参数和噪声统计的多传感器线性离散随机系统,应用递推辅助变量(RIV)算法和求解相关函数矩阵方程方法,得到模型参数估值器和噪声统计估值器。对带相关观测噪声和未知噪声统计系统,用经典Kalman滤波方法,基于Riccati方程,在按分量标量加权线性最小方差最优信息融合准则下,分别提出了自校正分量解耦信息融合Kalman和Wiener估值器。对带有未知模型参数和噪声统计的AR信号提出了自校正信息融合Wiener滤波器。用动态误差系统分析(Dynamic Error System Analysis)方法证明了自校正信息融合估值器的收敛性。几个跟踪系统的仿真例子说明了其正确性和有效性。
【Abstract】 One of the important methods of Multisensor information fusion is to detect the same target by multiple sensors, and then the local state estimates can be obtained ,under certain optimal fusion rules, by combining or weighting the local estimates, the optimal fusion estimates can be obtained, whose accuracy is higher than local estimates.Self-tuning information fusion filtering is used to deal with information fusion filtering problems for the multisensor systems with the unknown model parameters and noise statistics, it is a frontier feild between optimal information fusion filtering and system identification, so it has important theoretical and applied significance.For the multisensor systerms with unknown model parameters and noise statistics,using recursive instrumental variable (RIV) algorithm and solving correlation function matrix equations,the model parameters estimators and noise statistics estimators are obtained. For the multisensor systerms with correlated measurement noises and unknown noise statistics,using the classical Kalman filtering method, based on the Riccati equation, under the linear minimum variance optimal information fusion criterion weighted by scales for components, the self-tuning component decoupled information fusion Kalman and Wiener estimators are presented respectively. For the AR signals with unknown model parameters and noise statistics, a self-tuning information fusion Wiener filter is presented. The convergence of the self-tuning information fusion estimators is proved by the dynamic error system analysis(DESA) method. Several simulation examples of tracking system show their effectiveness.