Abstract:

The ducted fan micro aerial vehicle (MAV) can hover, take off and land vertically and fly forward with high security, compact structure and low noise. However, the flight dynamics of a ducted fan MAV is remarkably unsteady and nonlinear due to its big flight envelope, special aerodynamic configuration, low speed, small size, and complex flight environment. In view of this condition, a new control method of dynamic inversion with neural network is employed. A new flight control system of the ducted fan MAV is composed of dynamic inversion controller, neural network compensators, proportion-derivative (PD)compensators and pseudo-control compensators. Simulation results demonstrate that the adaptive inverse control system is much more stable and robust than the dynamic inversion-proportion-integral-derivative (PID)control system.