ROBOTIC UNMANNED AERIAL VEHICLE TRAJECTORY TRACKING CONTROL

Control of Unmanned Aerial Vehicles is a difficult task. One of the most challenging problems are nonlinear nature of the vehicle dynamics and the second is related to the nonminimum phase system. There are several methods to design nonlinear controller. One of them is so-called feedback linearization. Even, after successful design of nonlinear controller several important issues like nonminimum phase problem remain. The system with nonminimum phase dynamics needs to have outputs redefined. The output redefinition technique is used in a way such that the resulting system to be inverted is a minimum phase system. The complex calculation related to the system dynamics redefinition make a real-time computation very difficult. The real-time control system needs to be fast and reliable. In order to make a real-time control possible, a neural networks method has been developed and presented. The NARMA-L2 Neural Network is trained off-line to identify the forward dynamics of the UAV model with the redefined output, which is subsequently inverted to force the real output to approximately track a command input. Simulation results show that the proposed neural network method has overall good performance.