基于UD分解的改进型强跟踪滤波器

带多重次优渐消因子的强跟踪滤波器在计算舍入误差的影响下,估计状态的均方误差矩阵会失去非负定性,从而导致滤波发散的现象.在UD分解算法的基础上,提出了一种改进型带多重次优渐消因子的强跟踪滤波器.通过对引入渐消矩阵的一步预测均方误差矩阵进行UD分解,并将分解矩阵应用于滤波器量测信息的更新过程中,增强了算法的数值稳定性.此外,根据UD分解的结果给出了渐消矩阵的最小二乘解算方法,在提高渐消因子计算效率的同时摆脱了对先验知识的依赖性.仿真结果表明了算法的有效性.

Modified strong tracking filter based on UD decomposition

In the strong tracking filter with multiple suboptimal fading factors,the non-negative definiteness of the estimated state's mean square error matrix might be lost under the influence of the round-off error during calculations,thus resulting in the divergence of the filtered data.A modified strong tracking filter with multiple suboptimal fading factors is proposed based on the UD decomposition algorithm.By the UD decomposition of the one-step forecast mean square error matrix in which the fading matrix is introduced,the decomposition matrices are used in the process of updating the filter measurement information,then the stability of the algorithm's numerical value is intensified.Besides,a method of calculating the fading matrix is presented based on the least squares,not only the calculation efficiency of fading factors is improved but their dependence on the apriori knowledge is removed.Simulation results show the effectiveness of the algorithm.