Launcher servo-system model identified by an improved GA-BP NN

Referring to the random and nonlinear interference and the slow change of parameters of launcher servo-system, a new algorithm was put forward to identify launcher servo-system model by combining the characteristics of the genetic algorithm and improved BP algorithm. The principle was expounded, and the algorithm flow and formulas were presented. It overcomes the shortcoming of BP NN, such as the slow learning rate, and easy to converge to the local minima. The result of simulation shows that the algorithm has greatly improved the convergent accuracy and speed of NN, and gets a good identification result.