摘要
针对新息滤波交互式多模型(IF IMM)算法中切换过程模型概率滞后的问题,提出了模型概率转移矩阵马尔可夫参数自适应的新息滤波多模型算法(AM P-IF IMM),该方法采用后验信息修正不准确的先验信息,自适应的调整马尔可夫转移矩阵的参数.切换时刻较多地遗忘非匹配模型的信息,放大匹配模型的信息,在保证滤波精度的同时,大大提高了模型间切换速度.将该算法应用到CA,CV两模型组合导航系统取得了良好的效果.
The innovation filtering interacting multiple model estimator (IFIMM) performs more accurate result than conventional IMM, but it may be imposed to slow down the model switching. An adaptive Markov parameter IF-IMM algorithm (AMP-IFIMM) is proposed, in which the Markov transition probabilities can be modified adaptively during the process of filtering. By omitting the information of non-matching model and magnifying the matching model information simultaneously at the switching time, the algorithm can meet both the requirements of accuracy and switc- hing speed. The application of the algorithm on CV-CA two model integrated navigation system takes out reasonable resuit.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第3期521-524,共4页
Acta Electronica Sinica
基金
国防科工委基础科研项目-"微小型系统智能控制与信息处理技术"(No.J1600B001)
关键词
交互式多模型
新息滤波
马尔可夫链
组合导航
interacting multiple model
innovation filtering
Markov chain
integrated navigation