摘要
本文研究工程实践中应用Kalman滤波时出现的滤波异常现象。文中从滤波异常的检测出发,由此构成动力学噪声方差阵的自适应估计。对于常用的衰减记忆滤波,提出了衰减因子的自适应确定方法,对于未知的动力学噪声和观测噪声的方差阵,运用Bayes-ML估计方法给出了递推估算方案。对于构造自适应滤波的技术途径,提出了一些看法。
This paper discusses the exceptional phenomeina for using kalman filter in nsual engineering practice.At first,we start from detection of the anomaly of filtering to construct the algorithm for adaptive setimation.Then,for decay mamory filtering,We give an adaptive estimation of decay factor.About the dynamic model and observational model with unknown noise statistics,We design ML-estimation with recursive algorithm.Some approaches of constructing adaptive filtering are also considered at the same time.
出处
《国防科技大学学报》
EI
CAS
CSCD
北大核心
1994年第3期68-79,共12页
Journal of National University of Defense Technology