Comparison Between the System Identification and the Neural Network Methods in Identifying a Model Helicopter's Yaw Movement

The system identification and the neural network are often used to identify the micro helicopter's yaw models. In order to acquire a better method, models derived from the two methods are compared and analyzed in three aspects: data procession, model precision and the usability in the on-board control system. The results of the comparison suggest that it is a good way to use the neural network to pre-process the experiment data and use the system identification method to identify the model. In the comparison, a weighted criterion is also investigated to process the multi-experiment data and a neuron is used to determine the weight factors of the different experiment data.