一种基于最大似然概率准则的迭代DDF算法  

A Novel Iterated DDF Algorithm Based on Maximum Likelihood Probability

作　　者：辛勤[1] 吴顺华[1] 

机构地区：[1]国防科技大学电子科学与工程学院一所信号处理室,湖南长沙410073

出　　处：《信号处理》2010年第7期1033-1038,共6页Journal of Signal Processing

基　　金：武器装备预研基金资助项目(批准号:9140A20010107KG0163)

摘　　要：随着定位跟踪技术的不断发展,非线性滤波逐渐成为研究的重点和实现非线性定位跟踪的关键。DDF算法是一种基于stirling内插公式的非线性插值滤波算法,在系统的可观测度较低且测量误差通常较大时,其跟踪滤波的收敛速度、精度和稳定性都不高。在推导了最大似然概率迭代策略的基础上,提出基于最大似然准则的IDDF滤波算法。该方法迭代过程以似然概率增加为准则,改善了跟踪滤波精度和收敛速度。仿真实验表明,与EKF和DDF相比,IDDF具有更高的估计精度和更快的收敛速度。With the development of the location and tracking technology,the nonlinear filtering is becoming the research emphasis,which is the key to realize the nonlinear location and tracking. DDF is a nonlinear interpolation filtering algorithm based on stirling interpolation formula. When the system observability is weak and the observation error is large,the convergence speed,accuracy and stability of DDF are inferior. A novel iterated DDF （IDDF） is proposed by introducing the maximum likelihood probability iterated means into DDF. Since the likelihood probability is always increased in the iterated process of IDDF,its tracking accuracy and convergence speed are improved. Simulations are designed and carried out. The results indicate that,compared with EKF and DDF,higher accuracy of estimation and faster convergence are obtained using IDDF.

关 键 词：非线性滤波 差分滤波 迭代差分滤波 最大似然概率 

分 类 号：TN911.7[电子电信—通信与信息系统] TP273[电子电信—信息与通信工程]