|M.Sc Student||Nazarov Sergey|
|Subject||Modeling, Identification and Robust Control of a|
|Department||Department of Mechanical Engineering||Supervisor||Professor Leonid Mirkin|
|Full Thesis text|
This work addresses modeling and control of a 3DOF low-cost laboratory helicopter, controlled through adjusting the speed of its two propellers. The challenges are highly nonlinear dynamics, noisy sensors, impossibility to measure the angle / angular velocity of the propellers, unreliable components, electromagnetic coupling in control channels, and the instability of the open-loop system. A linear design model of the helicopter is derived by combining modeling from first principles, open- and closed-loop identification of system components, and linearization techniques. This results in a two-input (armature voltages of two DC motors) and three-output (pitch, roll, and yaw angles) system with non-negligible inter-channel couplings.
A two-loop controller design procedure to control the pitch and yaw angles is proposed. The inner loop is closed on the pitch and roll channels using the H-inf loop shaping design procedure. To this end, decoupling and PI weighting functions are used to render the design more transparent and to guarantee the presence of an integral action in the controller. The outer loop is then closed on the yaw channel using the roll angle as its control input. A nonlinear generator of reference signals for the yaw channel is also proposed. It generates a trajectory that reaches any required yaw set-point in minimal time under acceleration and velocity constraints. This reference signal is easy to tune and its use facilitates the avoidance of integrator windup phenomena. The designed controller is experimentally validated.