臺灣博碩士論文加值系統

在因應控制需求而進行建模時，找出受控體的數學模式是件甚具挑戰性的工作，於是各種方法紛紛出籠；不是將系統降階或將系統以線性化模型近似，就是將系統中某些被認為不重要的參數或影響予以忽略，以便簡化分析的複雜度。由於作了太多不真實的假設及簡化忽略，到最後所得到的數學模式即使可快速且精確地算出控制量，但跟實際上的物理系統可能出現相當大的差距，容易與實際脫節，以致於產生性能不佳的控制。
現代設備複雜性日益增加，其精確的數學模式更加難以獲得。傳統控制方法對控制系統進行分析和設計，其數學模式的求取是透過實驗的方式，以量測而得的頻率響應，求出控制系統的簡化模型，再由簡化模型的轉移函數來描述系統的動態特性。自從 Ziegler和 Nichols二位學者提出 PID控制器的調整公式以來，各種PID控制器的研究不斷產生，當受控體的數學模式限定在線性一階或二階系統時，PID控制器確實足以勝任。然而，PID控制器對於較高階、複雜的受控系統，仍採用二階簡化模型替代，以致於對其控制性能不彰或無法控制。另外，若受控體參數因環境或人為因素影響有所變動時，PID控制器的參數亦無法線上即時調整，使其強健性能受到影響。
鑒於上述缺失，本論文發展出智慧型控制中的可信度分配小腦模型控制器，以實現具即時調整參數能力之高性能PID控制器。所研究結果使用C++程式語言撰寫線上小腦模型演算法，配合Matlab/Simulink軟體模擬下列系統：感應電動機、飛行船、具有穩態誤差的二階系統及二階不穩定系統等。由模擬結果可知，本文所提之PID型小腦模型控制器能有效控制上列系統，且在性能上有顯著之改善。

As modeling the physical systems to meet the need of control purpose, finding mathematical models of these systems is a challenging work, and there are many approaches proposed to solve this problem. To reduce the complexity of analysis, model reduction and linearization techniques are frequently adopted. Also, insignificant parameters or inferences are sometimes neglected during the modeling process. Though using the forgoing obtained mathematical models can easily and quickly calculate the control force, in some circumstances, the control behavior of simulation is deviated from that of implementation because of too many unpractical assumptions and oversimplification. Consequently, the controller designed based on the mathematical model did not work properly.
Due to the increasing complexity of modern controlled objects, accurate mathematical models are getting difficult to obtain. Mathematical models used in conventional control system analysis and design are established by way of experiments. From the frequency responses of experiments, the simplified models of the control system are then built by zero-pole matching and then the transfer function of the derived model to can be used to describe the dynamic characteristics.
Motivated by Ziegler and Nichols PID tuning method, there are various PID controller design methods suggested. When the mathematical model is of order 1 or order 2, the designed PID controller performs well. However, for the high order systems, the performance of designed controller based on the second order system deteriorates and the closed-loop system is even out of control in the worst case. Moreover, when the parameters of plant fluctuate due to the environmental or artificial factors, PID parameters can not on-line and real-time self adjust. As a result, the robustness of closed-loop system is affected.
To overcome the aforementioned disadvantages and implement the high performance PID controller with real-time parameter tuning ability, this thesis proposes a new Credit Assignment Cerebellar Model Articulation Controller (CA-CMAC), which is widely investigated in the intelligent control field. The research result of this thesis, the on-line CA-CMAC algorithm, is programmed by C++ language and accompanied with Matlab/Simulink to simulate the following plants: induction motor, aircraft, second-order system with steady-state error, and second-order unstable system. From the simulation results, it is seen that the proposed PID CA-CMAC can effectively control these plants and the performances of dynamic behaviors are improved dramatically.

摘 要 i
ABSTRACT iii
誌 謝 v
目 錄 vi
圖目錄 viii
第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 2
1.3 研究範圍與限制 2
1.4 研究步驟 3
第二章 小腦模型控制器的理論基礎 5
2.1 控制系統的性能評價 5
2.2 小腦模型控制器理論架構 7
第三章 高性能小腦模型控制器設計 13
3.1 前言 13
3.2 傳統小腦模型控制器之工作原理 13
3.3 高性能小腦模型控制器之工作原理 15
3.4 結語 21
第四章 高性能PID型小腦模型控制器實例模擬 22
4.1 模擬一：感應電動機 23
4.2 模擬二：飛行船 26
4.3 模擬三：具穩態誤差的二階受控體 29
4.4 模擬四：具穩態誤差的二階受控體（欠阻尼） 33
4.5 模擬五：延遲受控體 36
4.6 模擬六：二階不穩定受控體（一） 39
4.7 模擬七：二階不穩定受控體（二） 42
第五章 結論與建議 46
5.1 結論 46
5.2 建議 46
參考文獻 48

[1]Tao Liu, Danying Gu, and Weidong Zhang, “A H-infinity Design Method of PID Controller for Second-Order Processes with Integrator and Time Delay,” Proceedings of 42nd IEEE Conference on Decision and Control, 2003, pp. 6044-6049.
[2]J. S. Albus, “Data Storage in the Cerebellar Model Articulation Controller (CMAC),” Transactions ASME, Journal of Dynamic System Measurement and Control, September 1975, pp. 228-233.
[3]J. S. Albus, “A New Approach to Manipulator Control: The Cerebellar Model Articulation Controller (CMAC),” Transactions ASME, Journal of Dynamic System Measurement and Control, September 1975, pp. 220-227.
[4]W. T. Miller, “Real-time Application of Neural Networks for Sensor-based Control of Robots with Vision,” IEEE Transactions on Systems, Man and Cybernetics, Vol. 19, 1989, pp. 825-831.
[5]C. C. Lin and F. C. Chen, “On a New Control Scheme, and Its Comparisons with the PID Controllers,” Proceedings of the American Control Conference, Arlington, June 25-27, 2001, pp. 769-774.
[6]J. Moody, “Fast Learning in Multi-resolution Hierarchies,” Proceedings of Advances in Neural Information Systems, D. S. Touretzky, Ed. Los Altos, CA: Morgan Kaufmann, vol. 1, 1989, pp. 29-39.
[7]M. Hormel, “A Self-organizing Associative Memory System for Control Applications,” Proceedings of Advances in Neural Information Systems, D. S. Touretzky, Ed. Los Altos, CA: Morgan Kaufmann, vol. 2, 1990, pp. 332-339.
[8]M. J. Carter, F. J. Rudolph, and A. J. Nucci, “Operational Fault Tolerance of CMAC Networks,” Proceedings of Advances in Neural Information Systems, D. S. Touretzky, Ed. Los Altos, CA: Morgan Kaufmann, vol. 2, 1990, pp. 340-347.
[9]L. G. Kraft and D. P. Campagna, “A Comparison between CMAC Neural Network Control and Two Traditional Adaptive Control Systems,” IEEE Control System Magazine, Apr. 1990, pp. 36-43.
[10]R. Smalz and M. Conrad, “Combining Evolution with Credit Apportionment: A New Learning Algorithm for Neural Nets,” Neural Networks, vol. 7, no. 2, 1994, pp. 341-351.
[11]Yiu-fai Wong and Athanasios Sideris, “Learning Convergence in Cerebellar Model Articulation Controller,” IEEE Transactions on Neural Networks, Vol. 3 No. 1, January 1992, pp. 115-121.
[12]W. T. Miller, F. H. Glanz and L. G.. Kraft, “CMAC: An Associative Neural Network Alternative to Backpropagation,” Proceedings of the IEEE, vol. 78, no. 10, 1990, pp. 1561-1567.
[13]Gene F. Franklin, J. David Powell and Abbas Emami-Naeini, Feedback Control of Dynamic Systems, Addison-Wesley Publishing Company, Inc., 1994.