| The main focus of this dissertation is the modeling and control of rigid and flexible manipulators. A modeling scheme for planar flexible multi-link manipulators is developed based on finite element and Lagrange methods. It is shown that this method has advantages over the assumed modes method. It is believed that the current model can provide more accurate prediction of the motion of flexible manipulators than any other available models. The dynamic characteristics of flexible manipulators are studied through modal analysis and numerical simulations. A number of key issues regarding the dynamics of flexible manipulators are discussed, e.g., natural frequencies and mode shapes, convergence of the modeling, and effects of configuration and loading.; A new control scheme, adaptive robust control, is proposed for control of rigid robot manipulators. This control scheme combines adaptive control and robust control in such a way that adaptive control compensates for the parameter uncertainties and robust control compensates for disturbances. Both trajectory tracking error and parameter estimation error are uniformly ultimately bounded in the presence of parameter uncertainties and unmodeled dynamics.; Singular perturbation theory combined with robust control is applied for the control of flexible manipulators. The dynamic model is reduced to two time scaled submodels by the application of the singular perturbation technique. The robust controller is applied for the slow submodel. The linear optimal controller (LQR) is used to control the fast submodel, e.g., vibration of the flexible links. First, it is shown that the the separate submodels can be stable with the developed control law. Then the stability of the complete system is analyzed using Lyapunov's second method. From the analysis, it is shown that, when the perturbation parameter is appropriate, the complete system is stable in the sense that it is uniformly ultimately bounded. Numerical simulations are provided to support the analysis. |
