Quadrotor Altitude Control using Recurrent Neural Network PID

The quadrotor is one type of Unmanned Aerial Vehicle (UAV) or unmanned flying vehicle. Quadrotor can be operated by a remote controller or autonomously. Quadrotor control is a challenging problem because it takes into account complex things such as parametric uncertainty, external disturbances, and so on. At the spatial level, three linear degrees of freedom along three axes and three degrees of freedom rotating along three axes are used for the control of a quadrotor. Conventional controls for quadrotors are widely used such as PID, state feedback, and so on. However, because the control is linear, non-linear control has begun to be developed. Some of these controls, for example, use a sliding mode control system, fuzzy methods, and controls by combining linear control with artificial intelligence. This paper will use PID control and an artificial neural network for the quadrotor direction control system. The results of this control test indicate that the combination of PID and RNN on the directional control shows a better response than conventional PID.