Online parameters identification of high speed train based on Gaussian Sum theory

In this paper, the relationship between traction, resistance, braking force, speed and acceleration is discussed. A nonlinear parametric state space model is established to describe the dynamic characteristics of running process of high speed train. Further, a filtering method based on Gaussian Sum theory and extended Kalman filter is proposed for estimating the states and parameters of nonlinear systems, and is applied to the high speed train model no matter it is affected by Gaussian or non-Gaussian noise. Firstly, the probability density function (PDF) of stochastic noise is approximated by a weighted sum of Gaussian PDFs with various means and variances. Then Bayesian theory and extended Kalman filter are combined to estimate the running states and model parameters online. Lastly, the running process of high speed train is simulated. The process is affected by uniformly distributed noise, and the running states and model parameters are estimated online by the proposed method. The simulation results show that the change of states and parameters can be effectively tracked, which demonstrates the effectiveness and feasibility of the proposed method.