基於模糊類神經網路與改良式通訊干擾估測器之時間延遲補償系統

A Time Delay Compensation System Based on Fuzzy Neural Networks and Improved Communication Disturbance Observer

簡達鴻(Da-Hong Jian)

指導教授 ： 李俊賢

國立臺北科技大學/電資學院/電機工程系研究所/碩士(2012年)

https://doi.org/10.6841/NTUT.2012.00470

摘要

近年來，由於網際網路的蓬勃發展，加上受控系統日趨複雜化，無論在工業以及學術上，傳統控制已無法滿足所需的控制性能，於是網路控制系統(networked control system, NCS)這方面的研究逐漸受到重視。然而在實際控制系統中，特別是透過網路來進行控制，時間延遲是一項急需解決的問題，過大的時間延遲會破壞整個系統的穩定性與性能指標，造成系統難以分析與設計。
因此，本論文提出一套時間延遲補償方法，以通訊干擾估測(communication disturbance observer, CDOB)架構為基礎，搭配模糊理論與類神經網路演算法，有效改善時間延遲對系統的影響。首先，利用模糊控制器取代傳統控制器進行設計，並與改良式CDOB架構作結合，針對CDOB架構中低通濾波器進行探討，隨著干擾訊號頻率的改變，模糊控制器可適應濾波器各種頻率之變化。系統建模部分則使用類神經網路之倒傳遞方法，透過不斷的訓練以及調整權重值，讓估測與實際的系統更為接近，使得建模的誤差降低，進而改善系統響應的穩態誤差。本文所提出的控制方法經由模擬結果得知，可以有效的應用在具有時間延遲之網路控制系統。

關鍵字

網路控制系統 ； 時間延遲補償 ； 通訊干擾估測 ； 模糊控制 ； 類神經網路

並列摘要

In recent years, due to the development of the Internet, the controlled systems
have become more complex. In industry and academia, traditional control has been
unable to satisfy the requirement of the control performance, so the research on
networked control system (NCS) is important increasingly. However, in the actual
control system, especially controlled over the Internet, the time delay is an
urgent problem. If the time delay is too long, the stability and performance of the
whole system will be destroyed, and then it is difficult to analyze and design the
systems.

Therefore, this paper proposes a set of time delay compensation method, based
on the architecture of communication disturbance observers (CDOB) with fuzzy theory and neural networks algorithm, and results show that it can improve the time delay system effectively. First, it replaces the traditional controller with fuzzy controller, combined with improved CDOB architecture, to study the effect of the low-pass filter in CDOB architecture. When the frequency of the disturbance signal changes, the fuzzy controller can be adapted to it. System modeling utilizes back propagation method in neural networks, by continuously training and the adjustment of weight value, the estimation is more close to the actual system, and therefore we can reduce the modeling error and improve the steady state error of the system response. From the simulation results, the proposed method in this paper can be effectively applied to the networked control systems with time delay.

並列關鍵字

NCS ； Time Delay Compensation ； CDOB ； Fuzzy Control ； Neural Networks

參考文獻

[1] A. R. Gupta and M. Y. Chow, “Networked control system: overview and research trends,” IEEE Trans. Industrial Electronics, vol. 57, no. 7, pp. 2527-2535, July 2010.

[4] B. Rahmani and A. H. D. Markazi, “Networked control of industrial automation

[7] C. L. Lai and P. L. Hsu, “Design the remote control system with the time-delay estimator and the adaptive Smith predictor,” IEEE Trans. Industrial Informatics, vol. 6, no. 1, pp. 73-80, Feb. 2010.

[9] T. Shi, S. Wang, H. Fang, and Z. Chen, “Fuzzy immune PI control of networked control system based on ADW time-delay prediction,” in Proc. Fuzzy Systems and Knowledge Discovery FSKD'08, Jinan, China, Oct. 2008, pp. 76-80.

[11] Z. Yusuf, Z. Janin, and M. Nasir Taib, “Application of fuzzy PD controller with Smith predictor for glycerin bleaching temperature control,” in Proc. Intelligent and Advanced Systems ICIAS’10, Kuala Lumpur, Malaysia, June 2010, pp. 1-5.