| Summary: | 碩士 === 國立成功大學 === 電機工程學系碩博士班 === 95 === Due to the fact that the bandwidth of optic fiber transmission is several hundred times greater than cable transmission and optical communication technology is well-developed and has advanced in recent years, the optic fiber communication system has become the mainstream of wired communication. However, the coupling procedure between optic fiber and the waveguide is very complex so that the coupling loss caused by disalignment is inevitable. This creates an obstacle to cost reduction. Hence the technological demand of auto-assembly and positioning control cannot be overlooked.
The advantages of the Piezoelectric actuator (PEA) include high resolution, compactness, low energy consumption, and rapid response. This thesis takes advantages of PEA to develop a vision-based optical fiber alignment platform. However, the tracking control accuracy of PEA is limited due to its inherent hysteretic nonlinearity and time varying parameters. In order to cope with this difficulty, a feedforward inverse hysteresis controller based on a B-spline approximation model is adopted to eliminate the hysteresis effects. In addition, a fuzzy logic controller is designed to reduce the error due to noise and model uncertainty. The accuracy and robustness of the servo control scheme are verified via a variety of reference inputs.
In addition, the vision latency problem makes the visual command unsmooth and the performance is degraded. In order to improve the performance of the system with visual feedback, the idea of multirate control was exploited to shorten the sampling period of the control system. Experimental results show that the proposed approach exhibits satisfactory performance.
|
|---|
