基于HMM模型的贝叶斯滤波方法

Bayesian Filter Approach Based on HMM

针对非线性动态系统估计问题,引入了一种不完全观测数据的统计模型——HMM模型,其双层随机结构较好地满足了实际估计的要求.在具体滤波过程中,将处理非线性问题常用的贝叶斯方法和HMM模型进行结合,通过预测和更新操作实现系统状态后验的连续递推估计,提高了运算速度.

About the problem of nonlinear dynamic system estimation, the authors introduce the HMM, a kind of statistical model based on incomplete data, whose two- double random structure can meet the need of solving actual estimation problems. In the flitting process, by combining bayesian filter with HMM, the estimation of the posterior density of the systemic state through the predicting arithmetic and the updating arithmetic can be implemented continuously. At the same time, the velocity of operation can also be improved.