Direct inverse control based on neural network for unmanned small helicopter attitude and altitude control

This paper describes the development of inner loop control of a small helicopter using artificial neural network. A Direct Inverse Control (DIC) based on Neural Network is proposed to maintain the attitude (roll, pitch, and yaw) position on the helicopter hovering state, for several different altitude values. The adopted neural networks learning method to train the network is the backpropagation algorithm. Simulations using real experimental helicopter hovering flight data were conducted to verify the performance of the proposed NN-DIC system. It was revealed that the simulated NN-DIC system can follow the hovering trajectory reference with very low error values.