Robust Flight Controller Design For Model-scale Helicopters Via Parametric Approach

The development of key technologies, such as sensors, microelectronics andcommunications, has led to the birth of unmanned aerial vehicles (UAVs) and hasfurther brought a huge progress in their development. Because rotorcraft UAVshave the great agility and maneuverability, they can autonomously completevarious missions in restricted or hazardous areas, such as combat gunfire support,reconnaissance and surveillance, inspection on power transmission, scientificinvestigation etc. And the research and development of rotorcraft UAVs havegained much attention in the academic and military communities worldwide asthey have potential applications in the military and civil areas. When model-scalehelicopters are equipped with necessary sensors, data processing units andwireless communication devices, they can be built into rotorcraft UAVs. To makethe model-scale helicopters as platforms for experiments and applications, it isurgent to enable them to hover and forward flight autonomously. In order todesign the helicopters’ autonomous flight control system, this thesis establishesthe complete mathematical model of model-scale helicopters, and design hoverand forward flight controllers via robust eigenstructure assignment.Firstly, we assume the rotor is perfectly rigid with no twist and can’t flap. Atthe same time, the rotation speed of main rotor is constant and model-scalehelicopters flight at low speed. Under these assumptions, this thesis presents themathematical model of model-scale helicopters. The basic idea of modeling is todecompose the entire helicopter system into the fuselage subsystem and forceand torque generation subsystem. Because fuselage kinematics and dynamicsequations are relatively simple, the thesis mainly considers the dynamics of themain rotor subsystem and its actuator agencies which include swashplate andflybar. We first give the expression of the cyclic servo inputs applied to flybarand main rotor by swashplate, and then formulate flybar’s flapping movementand the interaction between the main rotor and flybar. We eventually establish themathematical model of model-scale helicopters by letting all forces and torquesact on the fuselage. In order to simplify nonlinear aerodynamic characteristics ofthe flybar and main rotor, we also introduced two linear compensation factors inthe derivation. These two factors can be determined through identification.Secondly, with regard to the highly nonlinear and under-actuated property of the helicopter model and the coupling properties among channels, we make somenecessary simplifications to the nonlinear dynamics model of the helicopter inthis thesis. So we linearize the nonlinear model of helicopter around theoperating point. Based on the linear model, this paper analyzes the couplingproperties between the translational motions and the rotation motions in nearhovering flight. Analysis and simulations demonstrate that there exists strongcoupling properties between longitudinal translation and pitch movement, and thesimilar properties between lateral translation and roll movement can beestablished. But, in near hovering flight, the coupling properties between the yawmotion and lateral translation can be ignored. Accordingly, we eventuallydecompose the model into three subsystems, which includes the couplingsubsystem between roll motion and pitch motion, yaw motion subsystem andvertical translation subsystems.Finally, this thesis formulates a parametric design approach for model-scalehelicopters’ hover and forward flight robust controllers. We present theprocedures of robust eigenstructure assignment, and then design the hover andforward flight controllers. Numerical simulations verify the performance ofparametric controller and the robustness of the parameter perturbation, and showthat the parametric controller outperforms the LQR controller.