本論文針對吊車系統(Overhead Crane System)，提出基於滑動模式控制並利用視覺回授之定位與抗搖晃控制。由於吊車系統本身具有欠驅動(Underactuated)、複雜之非線性動態、難以模型化之參數誤差以及摩擦力等因素，造成控制之精確度與穩定性不足。另外，很難以較少的驅動器，同時實現軌跡追蹤控制快速定位及最小的負載晃動的性能。本文中提出一耦合台車動態(Trolley Dynamics)與負載搖晃動態(Load-swing Dynamics)之滑動平面(Sliding Surface)，利用滑動模式(Sliding-Mode)控制器的設計，將負載搖晃之動態引入阻尼(Damping)效果以消除擺動；另外，所提出的滑動模式控制並不需精確動態模型，即可提供足夠的強健性以克服複雜之非線性動態、負載變化與摩擦力等非理想因素，確保精確的定位達成期望之控制性能，並基於視覺回授架構所具有的量測、辨識、位置追蹤等特性，提升吊車絕對位置之參考輸入軌跡追蹤能力。模擬與實驗結果皆顯示出本文所提出之控制法則在貨物運送過程中，可確保漸進穩定之滑動平面、精確之台車位置追蹤以及優異之抗搖晃效果。 In this thesis, we present an anti-swing control scheme based on the sliding-mode control (SMC) for an overhead crane system using visual feedback. Since the overhead crane system is underactuated and highly nonlinear, the system performance is usually deteriorated by the system unmodelled-parameter errors and friction. On the other hand, to minimize load-swing angle and maximize the speed of load transfer are hard to consider and implement simultaneously by using fewer actuators. A sliding surface coupled with trolley and load-swing dynamics is designed to stabilize the load-swing dynamics by injecting the damping effect into the dynamics of the crane system. The proposed control law provides the robustness to unknown load variation and friction without an accurate dynamical model. Using visual feedback can improve the tracking capacity of absolute position. Simulations and Experiments show that the proposed control scheme ensures the asymptotically stability of sliding surface, precise trolley positioning, and good performance of the load-swing dynamics during the load transfer.
