说明：  提出了具有一般相关量测噪声的线性系统的平滑估计算法, 该算法是在系统正向和逆向滤 波估计结果的基础上,利用线性无偏最小方差估计获得的.由于量测噪声的相关性,使得其后验均 值不一定等于其先验均值,而它的后验均值又无法通过计算得到, 因而提出的算法是一个次优算 法.在正、 逆向滤波结果已知时,所提出的算法计算量小,易于实现.仿真实例说明,该算法的估计结 果要优于正、 逆向滤波估计结果,以及量测噪声不相关的Kalman 平滑估计结果
(Based on the forw ard and backw ard f i ltering est imates a smoothing algorithm is developed for linear systems w i th general correlated measurement noises by using the l inear unbiased minimum v ariance est imation formula. Because of the correlat ion of the measurment noises the posterior mean of the noise is not alw ays equal to its prior one and can
t be calculated. Hence, the proposed algorithm is subopt imal . When the forward and backward f il tering result s are know n, the proposed algorithm has low computional complex ity and can be real ized easi ly. T hrough a simulat ion example i t is indicated that the result of the proposed smoothing algorithm is bet ter than that of the forward, backward filter
ing or Kalman smoothing algorithm, where the measurement noises are assumed to be uncorrelated)