A maximum likelihood approach to state estimation of complex dynamical networks with unknown noisy transmission channel

In this paper, the problem of state estimation of complex dynamical network with unknown noisy transmission channel is considered. A likelihood function of the complex network is formulated. An expectation maximum (EM) algorithm is proposed to estimate the complex network´s state and the noise parameter simultaneously. The proposed method can obtain suboptimal estimate of the state and noise parameter. At each iteration of the EM algorithm, the complex network´s state is estimated by extend Kalman filter in the E-step, while the noise parameter is updated in the M-step. Computer simulation results verify the effectiveness of the proposed method.