High performance mechtronics servo system identification toolkit design

Abstract

Mechatronics servo control system in hard disk drive (HDD) plays an important role in detecting and predicting the errors where system identification parameters are crucial in the process. It has been debated that system identification data at high frequency range is always inaccurate. This project aims to create a GUI interface which automatically generates the transfer function accurately from the frequency response with minimal tuning. In addition, this report analyzes the different identification method and auto weighting algorithm.

Least square estimation and MATLAB transfer function identification were implemented in the mechatronics servo system identification toolkit with optimization algorithm. Results will be observed and compared in graphical approaches. The graph fitting gap and the difference in between the data points indicates the quality of the identification method. Besides, auto weighting algorithm is implemented to fine tune the results.

Both identification methods were observed to provide significant precise results in high frequency range under the different plant model. On the other hand, results indicated that the auto weighting algorithm aided the identification methods in optimization process.

To improve on results, more data points and plant model is to be tested to ensure the stability of the toolkit. As a conclusion, the mechatronics servo system identification toolkit provided a framework and platform for the control system identification process.