Time Series Model Identification And Its Application In The Missile Modeling

Time series models are widely used to analysis and forecast of the development of motion law of various system. In order to ensure the accuracy of the forecast, time series model is significantly established. In other words, the model parameters are accurately identified. The all parameters of time series model are identified, and time series model identification is applied to the missile model in this paper. The traditional problems of the missile body power complex coefficient and the large amount of calculation are solved.Firstly, the theory of time series model was introduced, and the least squares based iterative identification algorithm, gradient based iterative identification algorithm, multi-innovation stochastic gradient identification algorithm, multi-innovation least squares identification algorithm and maximum likelihood identification algorithm in the application of auto-regressive moving average model parameter estimation were studied. Through the simulation examples show the effectiveness of the algorithms and parameter estimation precision. Secondly, missile dynamics and kinematics model were established. And the differential equation of linearization method was used to linearize the missile longitudinal channel model. The linearized model transformed to transfer function of the longitudinal channel by the Laplace, including the transfer function of pitching angle and elevator angle, the transfer function of attack angle and elevator angle and the transfer function of trajectory inclination angle and elevator angle. Finally, the transfer function of attack angle and elevator angle was selected as the structure of identification model, the transfer function was discretized by using backward difference transformation method, then the auto-regressive moving average model was gotten. Using the least squares based iterative identification algorithm and maximum likelihood identification algorithm is to respectively identify the parameters of the model,and then to compare the two results of identification. It was found that the least squares based iterative identification algorithm was more suitable for the parameter estimation of the model. The stability analysis result verified the reliability of the model.The results of simulation show that:1, The least squares based iterative identification algorithm is more suitable for missile longitudinal channel model parameters estimation;2, Using identification algorithm of time series model is to estimate the parameters of the missile longitudinal channel model, which the estimated parameter values are close to the true value, and the errors of estimate values of parameter are within the scope of the permission. As a result, the method of missile longitudinal channel model parameter estimation is realized and reliable.