摘要
为解决大系统中应用整体式Kalman滤波遇到的非实时性难题,本文提出一种能保持最优的分布式Kalman滤波计算方法——逐次正交化法。这是一种适合多微机并行计算的最优算法,确保滤波的“实时性”,优于一般次优分布式Kalman滤波器。
A calculated method of successive orthogonalization procedure which can maintain optimal Distributed Kalman Filtration is presented in this paper. The difficult problem of non-real-time encountered in global Kalman Filtration used in large systems may be solved. This optimal algorithm is suitable for parallel processing on multiprocessor. Using this procedure, the calculation time may be shortened and the real-time of filtering can be ensured. This method is obviously superior to the general sub-optimal Distributed Kalman Filter.
关键词
卡尔曼滤波
最优化
分布式
正交化
Kalman filtering, large-scale system, optimization, distribution, orthogonalization
