Neural Network-Based Optimal Control of an Unmanned Helicopter

Unmanned helicopters are self-propelled and autonomously controlled rotorcraft that are capable of independent flight. Because of their versatility and maneuverability, these unmanned helicopters are indispensable for both civilian and military applications, where human intervention is restricted. For control of a helicopter [1], it is necessary to produce moments and forces on the aircraft with two goals: first, to position the helicopter in equilibrium such that the the desired trim state is achieved; and second, to control the helicopter’s velocity, position, and orientation such that it tracks a desired trajectory with minimal error. The dynamics of the unmanned helicopter are not only nonlinear but are also coupled with each other and underactuated, which makes control difficult. In other words, the helicopter has six degrees of freedom (DOF), which must be controlled with only four control inputs—thrust and three rotational torques. To solve the problem of controlling a rotary-wing unmanned aerial vehicle (UAV), several techniques have beenproposed [1–9]. The fullmodel of the helicopter is quite complex when the main rotor and tail rotor and effects of drag, actuator dynamics, and other disturbances are taken into consideration. Based onNewton– Euler equations, a dynamicmodel has been derived [1] considering the helicopter as a rigid body with input forces and torques applied to the center of mass. Also, it has been shown [1] that the multivariable nonlinear helicopter model cannot be converted into a controllable linear system via exact state-space linearization. In addition, for certain output functions, exact input-output linearization results in