| Cascade nonlinearity, e.g., friction, dead-zone, back-lash, etc., is one of the major obstacles in high precision motion control. Since the exact parameters of these nonlinear elements (e.g., corner points of hard discontinuities) cannot be easily identified, it is difficult to achieve high control performance.; This thesis reports our research work on studying the passivity properties of DC motor systems with cascade nonlinearity (represented in a Hammerstein model). It has been shown that Hammerstein plants are passive provided that the derived conditions are satisfied. In other words, the whole motion system can then be considered as a passive plant and passivity theorem can be adopted to govern system stability.; Identification of the system parameters plays an important role in high-precision motor control. If the nonlinearity and the dynamics of the motor system can be accurately identified, compensators or inverse controllers can be designed to counteract the nonlinearities. As a result, we will make use of the passivity properties of the plant to derive a stable and converging identification scheme to obtain system information. Comparison has also been performed to show the effectiveness of this approach against the commonly used iterative or IV identification approaches.; Based on the passivity framework, direct and indirect control schemes will be adopted to construct a stable adaptive system. Adaptive rules are designed such that the controller parameters could be optimized without sacrificing stability. Besides, it has been shown that intelligent control (e.g., fuzzy logic, neural networks) can be comfortably fit in without any structural changes. Simulation results illustrate that stability can be guaranteed with minimal system information. |
