V. Puttige and S. Anavatti (Australia)
Neural networks, control, nonlinear systems, aircraft, model.
In this paper a novel dual neural network based technique to control the pitch angle of an aircraft is presented. To control the nonlinear aerodynamics, two neural networks, one as an online trained controller and another as an in ternal model of the system, are used together. Numerical simulation results are validated using the real-time Hard ware in the Loop (HIL) simulation. Stability analysis of the controller shows under given boundary conditions the controller satisfies Lyapunov’s stability criteria. The neural networks model and controller are based on the Autore gressive architecture (ARX).
Important Links:
Go Back
