自動化設備係由伺服馬達連接不同機構負載所構成，因此設備之性能與伺服馬達響應息息相關，良好的伺服響應需要有良好的伺服運動控制器增益調校，而傳統的增益調整方法有計算困難、耗時與模型誤差等相關問題。本研究提出一種結合AI以及最佳化演算法之智能調機方法，透過蒐集伺服系統在不同控制器增益設計下之響應特徵規格，以AI演算法(LightGBM)建立響應特徵預測模型，再以最佳化演算法(PSO)最佳化控制器增益，其能針對伺服系統之特性求取最佳增益，又不需複雜的物理建模過程以及數值運算，且藉由最佳化的權重調整，可令調機之響應結果更容易貼合實際應用需求。
本文共分為四個部分，第一部分為介紹伺服馬達控制架構，以及各個控制器增益對於馬達響應之影響；第二部分則介紹傳統調機方法的問題，並且提出智能調機方法以改善傳統調機流程；第三部分為介紹所使用之LightGBM模型與粒子群演算法；第四部分則實際應用智能方法於伺服系統，證實此方法的有效性。

Automation equipment is usually composed of a servo motor connected to various mechanisms. The performance of the equipment is therefore closely related to the response of the servo motor. Servo motors are controlled by motion controllers, thus in order for the equipment to achieve optimal performance, the motion controller gains must be tuned appropriately. Previous research related to gain tuning is difficult to implement due to the inherent difficulty in system modeling and mathematical calculations, and traditional tuning methods are time-consuming and tedious.
This thesis presents a gain tuning method, “Smart Tuning”, which combines artificial intelligence (LightGBM) with an optimization algorithm (PSO). The LightGBM model predicts the response characteristics of a servo system under different controller gains, and the optimal controller gains can then be obtained through PSO. A fitness function, which is used to evaluate the motion controller performance is also formulated by combining several position response characteristics such as percentage of overshoot and settling time etc., and each characteristic is normalized to standardize the different units and scales. The experimental results show that appropriate controller gains can be obtained through the proposed tuning method.

[1]台達電子工業股份有限公司，ASDA-Soft軟體操作手冊，2013。
[2]台灣安川電機股份有限公司，Σ-7系列用戶手冊，2007。
[3]J.G. Ziegler and N. B. Nichols, "Optimum Settings for Automatic Controllers", Trans. ASME 64, pp. 759-768, 1942.
[4]H. Håkan, G. Svante, and G. Michel, "A Convergent Iterative Restricted Complexity Control Design Scheme, ", Proceedings of the IEEE Conference on Decision and Control, vol.2, pp. 1735 - 1740, 1995.
[5]J. C. Basilio and S. R. Matos, "Design of PI and PID controllers with transient performance specification", IEEE Transactions on Education, vol. 45, no. 4, pp. 364-370, 2002.
[6]M.C. Campi, A. Lecchini, and S.M. Savaresi, "Virtual reference feedback tuning: a direct method for the design of feedback controllers", Automatica, vol. 38, pp. 1337-1346, 2002.
[7]K. Hayashi and T. Yamamoto, "Design of a one-shoot tuning PID controller," IEEE International Conference on Control Applications, pp. 712-717, 2011.
[8]M. A. Siddiqui, S. H. Laskar, M. N. Anwar and M. Naseem, "A Model-Free PI/PID Controller based on Direct Synthesis Approach to achieve Disturbance Rejection," Annual Conference of the IEEE Industrial Electronics Society, pp. 207-212, 2019.
[9]J. B. He, Q. G. Wang, T. H. Lee, "PI/PID controller tuning via LQR approach", Chemical Engineering Science, vol. 55, no. 13, pp. 2429-2439, 2000.
[10]W. Tan, "Unified Tuning of PID Load Frequency Controller for Power Systems via IMC," in IEEE Transactions on Power Systems, vol. 25, no. 1, pp. 341-350, 2010.

[11]Z. L. Gaing, "A particle swarm optimization approach for optimum design of PID controller in AVR system," in IEEE Transactions on Energy Conversion, vol. 19, no. 2, pp. 384-391, 2004.
[12]S. M. GirirajKumar, Deepak Jayaraj, Anoop. R. Kishan, "PSO based Tuning of a PID Controller for a High Performance Drilling”, International Journal of Computer Applications, vol. 1, no. 19, 2010.
[13]C. Damour, M. Benne, C. Lebreton, J. Deseure, amd B. G. Perez, "Real-time implementation of a neural model-based self-tuning PID strategy for oxygen stoichiometry control in PEM fuel cell", International Journal of Hydrogen Energy, vol. 39, no. 24, pp. 12819-12825, 2014.
[14]蔡清雄，交流伺服原理及調機方法，台達電子工業股份有限公司，2007，
[Availible]: https://reurl.cc/Mvzxjn
[15]台達電子股份有限公司，台達廣用多機能型伺服驅動器ASDA-A2R系列應用技術手冊，2014。
[16]M. A. Sahib, B. S. Ahmed, "A new multiobjective performance criterion used in PID tuning optimization algorithms", Journal of Advanced Research, vol.7, no. 1, pp. 125-134, 2016.
[17]Y. Ju, G. Sun, Q. Chen, M. Zhang, H. Zhu and M. U. Rehman, "A Model Combining Convolutional Neural Network and LightGBM Algorithm for Ultra-Short-Term Wind Power Forecasting," in IEEE Access, vol. 7, pp. 28309-28318, 2019.
[18]G. Ke, Q. Meng, T. Finley, T Wang , W. Chen , W. Ma , Q. Ye , and T. Y. Liu, "LightGBM: A Highly Efficient Gradient Boosting Decision Tree", Advances in Neural Information Processing Systems 30, 2017.
[19]LightGBM 2.3.2 documentation,
[Available]: https://lightgbm.readthedocs.io/en/latest/Parameters.html
[20]C. Ou, W. Lin, "Comparison between PSO and GA for Parameters Optimization of PID Controller," International Conference on Mechatronics and Automation, pp. 2471-2475, 2006.
[21]台達電子股份有限公司，台達交流伺服驅動器ASDA-A2系列，
[Available]:http://www.shinwe.com.tw/wpcontent/uploads/2018/06/ASDA-A2.pdf
[22]J. Kennedy and R. Eberhart, "Particle swarm optimization," Proceedings of ICNN'95 - International Conference on Neural Networks, pp. 1942-1948 vol.4, 1995.