Paper accepted in AIAA SciTech
Table of contents
Link to the paper
Abstract
An autonomous unmanned aerial vehicle can be influenced by possibly unknown model uncertainties and exogenous perturbations. The primary sources of model uncertainties in an autonomous vehicle are noise and time delay in sensor measurements; however, an aerial vehicle is also exposed to exogenous perturbations due to wind gusts and ground effects (especially in the case of a quadrotor). This study proposes a novel control strategy for the soft landing of a quadrotor onto a textured stationary landing platform without any prior information on the model uncertainties and exogenous perturbations present in the system dynamics. The proposed controller employs the divergence of the optical flow field with the fundamental purpose of executing a constant flow divergence landing maneuver while driving the vertical velocity of the quadrotor to zero. The output error is subjected to prescribed performance constraints to achieve the desired transient and steady-state behaviour of the proposed controller. The design of this controller is based on proportional-integral (PI) sliding mode and a state-dependent control law that synthesizes continuous control actions to ensure finite-time convergence of the output error to a uniform ultimate bound. The critical control law is based on adaptive first-order sliding mode, where the control gains are designed to adapt to the unknown uncertainties and perturbations online. This control law also helps in attaining the finite-time convergence to a prespecified bound that is independent of the size of the uncertainties and perturbations. Subsequently, Lyapunov analysis is performed to demonstrate the global stability of the proposed control strategy. Numerical simulation results are presented to exhibit the effectiveness and robustness of the proposed control approach for different initial conditions in the presence of uncertainties and perturbations. Lastly, a performance comparison analysis is conducted with a fixed-gain PI controller to demonstrate the superiority of the proposed adaptive gain controller.
Video Presentation
View Video Presentation: link