Path Following Control for Multiple Quadrotors Carrying A Rigid-body Slung Payload

Novel robust path-following flight controllers for quadrotors carrying a tethered payload are extensively studied from the perspective of dynamic modelling, control design, and experimental verification. By using multiple quadrotors to cooperatively carry a payload, their payload capacities can be significantly boosted. The number of vehicles can be adjusted according to the weight of the payload, resulting in a flexible and efficient use of drone resources. The presented model development starts from a single quadrotor with a point-mass payload to multiple quadrotors with a rigid-body payload. The payload is towed by quadrotors with cables. The systems are decomposed into the payload subsystem and the quadrotor attitude subsystem by assuming the cable is tethered at the center of mass of each quadrotor. The controller designs are then developed for a single quadrotor with a point-mass payload, followed by controller of multiple quadrotors with a rigid-body payload. Both controllers resemble a cascade form in structure. The outer loop offers a robust path-following controller that stabilizes the payload subsystem by assuming the lift vector of each quadrotor can point instantaneously to a given direction. An uncertainty and disturbance estimator is designed to estimate and eliminate the lumped disturbances caused by exogenous wind and parameter imperfection. The inner loop, on the other hand, is an attitude tracker implemented on each quadrotor to follow a reference attitude generated by the outer-loop controller. The overall stability of the complete system is proven using the Lyapunov method and the Reduction Theorem. Aside from the analytical control law, a model predictive controller (MPC) method is also studied and implemented on quadrotor for cooperative slung payload delivery. The MPC method utilizes the equivalent damping force from the previous controller as the baseline stabilizing inner loop. The linearized closed-loop model is then calculated. Finally, the optimum controller is calculated after a prediction horizon and a cost function are defined. The MPC scheme achieves better performance and requires less parameter tuning. Extensive simulations and experiments show that the controller designs are capable of stabilizing the payload under model imperfection and exogenous disturbances simultaneously.Ph.D