Neural network model identification for actuators in a flight control systems

The precision system identification models is very important to design and analysis for complex control systems. A model of the actuators in a flight control systems is identified by using BP neural network. An adaptive BP neural network learning algorithm is proposed in this paper, which uses adaptive learn rate and steepest gradient optimization algorithm to train the weights. The improved algorithm is applied to identify the nonlinear actuators. The simulation results show that the performance of the neural network is improved effectively and the output of systems can be identified accurately by using the improved method.