Visual Servo Based Tracking and Balance Control of the Ball and Plate Systems

碩士 === 國立成功大學 === 工程科學系碩博士班 === 97 === In this thesis, a visual-servo and digital signal processor based ball and plate system is constructed. This system consists of a mechanism, two motor actuators, an image sensor, a DSP-based control card, the relevant peripheral interface circuits, and the soft...

Full description

Bibliographic Details
Main Authors: Li-Ming Chu, 朱立銘
Other Authors: Ming-Tzu Ho
Format: Others
Language:zh-TW
Published: 2009
Online Access:http://ndltd.ncl.edu.tw/handle/13846518272458933502
Description
Summary:碩士 === 國立成功大學 === 工程科學系碩博士班 === 97 === In this thesis, a visual-servo and digital signal processor based ball and plate system is constructed. This system consists of a mechanism, two motor actuators, an image sensor, a DSP-based control card, the relevant peripheral interface circuits, and the software programs. The ball and plate system is a highly nonlinear system. Due to existence of the centrifugal force, the system relative degree is not well defined. Moreover, the centrifugal force provides a strong positive feedback and easily leads to the peaking phenomenon. In this thesis, the decoupling method is used to obtain two independent ball and beam systems. The backstepping control techniques and approximate feedback linearization are then used to design the controllers. Backstepping control design is a recursive procedure that interlaces the choice of a Lyapunov function to discriminate stability with design of feedback control. The approximate feedback linearization technique regards some nonlinear terms as very small disturbances and neglects them. The original nonlinear system is then tranformed into an approximate input-output system. It is shown that approximate feedback linearization not only reaches the excellent performance, but also has less complexity in controller design and implementation. So that we can re-tune each parameter of the controller much easier. It is greatly helpful when we proceed the experiments. Both of the controllers can achieve stabilization, tracking and robust control under restrictive conditions. Moreover, they can reduce the effect of centrifugal force on the system, and make the system reach asymptotic stability.