Adaptive Neural Network Control of a Helicopter in Vertical Flight

In this paper we address the application of adaptive neural networks in the flight control of a miniature helicopter during hover. Neural networks have received a great deal of attention lately, especially in the areas of pattern recognition and control of unknown systems. In this paper we will discuss how to setup and update a two-layer neural network. We will also present an on-line control system which can be implemented using neural networks. Finally, we will show how neural networks can be used to improve the controlled response of a miniature helicopter in vertical flight by learning and adapting to changes in the plant on-line. Experimental helicopter flight control results are also given in this paper.