信号通路模型参数优化的非线性滤波方法

提出采用非线性滤波方法——平方根UKF估计信号通路模型的未知参数,利用协方差平方根代替协方差参加递推运算,保证了滤波算法的数值稳定性.以肿瘤坏死因子诱导的核转录因子κB信号转导网络为例,利用平方根UKF对系统模型的未知参数进行辨识.仿真结果表明,该非线性滤波方法能够从噪声数据中提取有效信号,提高了参数估计的精度,为解决复杂信号通路参数辨识中的不确定性问题提供了可靠的方法.

Nonlinear Filter for Parameters Optimization of Signal Pathways Model

A new nonlinear filter algorithm based on square root unscented Kaman filter(SRUKF) is presented and used to perform parameter identification.SRUKF takes covariance square root matrix instead of covariance one in filter iteration,which effectively avoided filtering divergence,and meanwhile improved the convergence velocity and stability of the algorithm.Using TNFα induced NF-κB signalling transduction pathways as an example,this method was applied to estimate the unknown parameters of the system model.Simulation results show that SRUKF is a more ideal nonlinear filtering method for estimating the unknown parameters under the disturbs of the noise and it provides an efficient theoretical method for solving the parameters' uncertainty effects of complex signal pathways.