臺灣博碩士論文加值系統

依據統計數據指出，大部分的空難是人為因素所造成的，且近半數的空難發生在民航機降落之際。此外，剪風與亂流也是造成空難的主要原因，而一般的飛行控制器，因充分考慮到飛機的控制性與乘客的舒適性，使反應變的頓化，對於像剪風等激烈的外在狀況變化，很難做迅速的反應。所以設計一智慧型的自動著陸系統取代目前傳統式控制方法的自動著陸系統有其必要性，此自動著陸系統可使飛行更為簡單，飛航安全更周全。本文之智慧型控制器採用模糊類神經網路結構，簡單來說，模糊類神經網路即是一個具有模糊推論能力之類神經網路，它結合了類神經網路與模糊邏輯兩者之優點。利用網路結構的鍵結值來調整歸屬函數的形狀及位置，使其具有適應性的能力，不用像傳統的模糊系統，根據嘗試錯誤法來調整歸屬函數，使系統最後的表現達到最佳化。也藉此決定了網路的結構大小，免去尋找最佳隱藏層數及神經元個數之困擾。學習模式採用倒傳遞演算法時，因為此學習過程並非動態行為，其結果可能不是很好，因此本論文主要提出以時序性倒傳遞演算法的模糊類神經網路控制器，並結合線性化反飛機模組，應用於飛機自動著陸系統，以期設計一智慧型的自動著陸系統，使其具有更優異的適應能力。最後，由電腦模擬結果顯示，應用傳統倒傳遞演算法訓練過的網路，做為飛機自動降落系統的控制器時，只能對 PI -controller做一映射的動作。應用時序性倒傳遞演算法訓練過的模糊類神經網路控制器於飛機的自動降落系統上，則可擴大外在環境之可控範圍，使飛機能承受較大的擾流並安全降落。

Fuzzy neural networks have been applied to flight control for their better adaptability and robustness for unmodeled systems and hardware implementation capability. Fuzzy neural networks can increase the flight controller’s adaptation to different environments. Currently, most of the improvements in the Automatic Landing System (ALS) have involved the guidance instruments. Using improved calculation methods and highly accurate instruments, these systems provide more accurate flight data to the ALS to make the landing smoother. However, these researches do not include weather factors such as wind shear. There have not been many researches on the problem of intelligent landing control. The purpose of this thesis is to apply a fuzzy neural network controller to an aircraft automatic landing system. In this research, we investigate a fuzzy neural network system that combines the advantages of the fuzzy logic and neural network systems. The Backpropagation Through Time algorithm is implemented into the network learning process. A complete landing phase is divided into several stages (intervals). Each stage uses the same fuzzy neural network controller. The Linearized Inverse Aircraft Model calculates the error signals that will be used to back propagate through the controller to obtain the weight changes in each stage. The error continues to be back propagated through all of the stages and weight changes for the controller are computed at each stage. The weight changes from all obtained stages from the delta learning rule are added together for the overall update. Simulation results show that the trained controller can guide the aircraft to land safely through wind disturbances and successfully expand the controllable environment in severe wind disturbances.

摘要 (中文) i
摘要 (英文) ii
誌謝 iii
目錄 iv
圖目錄 vi
表目錄 xii
第一章 導論 1
1.1 研究動機 1
1.2 文獻回顧 2
1.3 論文貢獻 4
1.4 論文大綱 4
第二章 飛行分析 5
2.1 飛行著陸分析 5
2.2 飛機線性動態方程式 8
2.3 安全降落的義 11
2.4 風擾數學式 12
2.4.1 亂流 12
2.4.2 剪風 14
第三章 倒傳遞模糊類神經控制 17
3.1 模糊邏輯 17
3.1.1 歸屬函數 17
3.1.2 模糊系統 18
3.2 類神經網路 22
3.2.1 類神經網路模型 22
3.3 模糊類神經網路 24
3.3.1 倒傳遞學習法則 29
3.4 模糊類神經控制器 32
3.5 模擬結果 35
3.5.1 於亂流中著陸控制 35
3.5.2 於剪風中著陸控制 46
3.6 結果討論 59
第四章 時序性倒傳遞模糊類神經控制 60
4.1 時序性倒傳遞演算法介紹 60
4.1.1 規則網路與其梯度之計算 60
4.1.2 循環式網路之時序擴展 61
4.1.3 時序性倒傳遞演算法 62
4.2 控制方法 65
4.3 線性化反飛機模組 67
4.4 模擬結果 69
4.4.1 於亂流中著陸控制 69
4.4.2 於剪風中著陸控制 78
第五章 結論 85
參考文獻 87

[1] A.J. Calise and R.T. Rysdyk, “Nonlinear Adaptive Flight Control Using Neural Networks,” IEEE Control Systems Magazine, vol. 18, no. 6, pp. 14-25, 1998.
[2] P.J. Werbos, “Neural Networks and Flight Control: Overview of Capabilities and Emerging Applications,” Proc. of the AIAA Guidance, Navigation, and Control Conf., pp. 912-919, 1995.
[3] M.R. Napolitano and M. Kincheloe, “On-Line Linearning Neural-Network Controler for Autopilot Systems,” Journal of Guidance, Control and Dynamics, vol. 18, no. 5, pp. 1008-1015, 1995.
[4] 謝成德, ”Gain Scheduling of Flight Control System with Fuzzy Logic Model,” M.S. thesis, 交通大學, 1996.
[5] 陳柏熹, “The Application of Neural Network on Longitudinal Flight Control System Design ,” M.S. thesis, 中興大學, 1994.
[6] 譚俊豪, ”模糊飛行控制系統的發展及應用於無人飛機之研究,” 國科會研究報告, NSC82-042-E006-148, 1993.
[7] Federal Aviation Administration, “Automatic Landing System,” AC 20-57A, Jan. 1971.
[8] B. Etkin, “Effect of Wind Gradient on Glide and Climb,” Journal of the Aeronautical Science, vol. 14, no. 6, 1947.
[9] C.C. Jorgensen and C. Schley, “A Neural Network Baseline Problem for Control of Aircraft Flare and Touchdown,” Neural Networks for Control, pp. 403-425, 1991.
[10] L.A. Zadeh, “Fuzzy Sets”, Inform. Control, vol. 8, pp. 338-353, 1965.
[11] 蔡宜成, 類神經網路模式應用與實作, 儒林圖書有限公司, 2000年4月七 版.
[12] R.P. Lippmann, “An Introduction to Computing with Neural Nets”, IEEE ASSP Magazine, pp. 4-17, April 1987.
[13] 陳燕慶, 鹿浩, 神經網路理論及其在控制工程中的應用, 儒林圖書有限公司, 1992年6月初版.
[14] J.G. Kuschewski and S.H. Zak “Application of Feedforward Neural Network to Dynamical System Identification and Control” IEEE Transactions on Control System Technology, vol. 1. no. 1, pp. 37-49, March 1993.
[15] M.S. Lan, “Adaptive Control of Unknown Dynamical System via Neural Network Approach,” Proc. American Control Conf., pp. 910-915, 1991.
[16] W.T. Miller, R.S. Sutton, and P.J. Werbos, Neural Networks for Control, the MIT Press, 1991.
[17] 陳士傑, “Fuzzy modelling using hybrid neural networks,” M.S. thesis,中央大學, 1995.
[18] 李慶鴻, “Analysis of Fuzzy Neural Networks and Its Applications,” M.S. thesis,交通大學, 1999.
[19] 林瑞杰, “Additive Gaussian Membership Functions in Fuzzy Neural Network,” M.S. thesis, 交通大學, 1999.
[20] R.C. Nelson, Flight Stability and Automatic Control, 4th edition, pp.314-315, Mcgraw-Hill International Editions, 1998.
[21] S. Asai, et al., “Development of Flight Control System for Automatic Landing Flight Experiment,” Mitsubishi Heavy Industries Technical Review, vol. 34, no. 3, 1997.
[22] D. Mc. Ruer, I. Ashkenas, and D. Graham, Aircraft dynamics and automatic control, Princeton University Press, Princeton, NJ, 1973.
[23] Y.H. Chen and S. Pandey, “Robust Control Strategy for Take-Off Performance in a Windshear,” Optimal Control Applications and Method, vol. 10, no. 1, pp. 65-79, 1989.
[24] C.T. Lin and C.S.G. Lee, Neural Fuzzy Systems: A Neuro-Fuzzy Synergism to Intelligent Systems, Prentice-Hall International, Inc., 1996.
[25] C.T. Lin and C.S.G. Lee, ”Neural-network-based Fuzzy Logic Control and Decision System,” IEEE Trans. On Computers, vol. 40, no. 12, pp. 1320-1336, 1991.
[26] W.S. McCulloch, and W. Pitts, “A logical calculus of the ideas immanent in nervous activity,” Bulletin of Mathematical Biophysics, vol. 5, pp. 115-133. 1943.
[27] F. Rosenblatt, “The perceptron：A probabilistic model for information storage organization in the brain,” Psychological Review, vol. 65, pp. 386-408, 1958.
[28] J.J. Hopfield, “Neural networks and physical systems with emergent collective computational abilities,” Proceedings of the National Academy of Sciences, USA, vol. 79, pp. 2554-2558. 1982.
[29] D.E. Rumelhart, G.E. Hinton, and R.J. Williams, “Learning representations by back-propagating errors,” Nature, vol. 323, pp. 533-536, 1986.
[30] P.J. Werbos, “Backpropagation Through Time: What It Does and How to Do It,” Proceedings of the IEEE, vol. 78, no. 10, pp. 886-901, September 1990.
[31] D.H. Nguyen and b. Widrow, “Neural Networks for Self-Learning Control” Systems Magazine, pp. 18-23, April 1990.
[32] R. Bulirsch, F. Montrone, and H. Pesch, “Abort Landing in the Presence of Windshear as a Minimax Optimal Control Problem,” Journal of Optimization Theory and Applications, vol. 70, no. 1, pp.1-23, 1991.
[33] J.G. Juang and C.S. Lin, “Giat synthesis of a biped robot using backpropagation through time algorithm,” Proc. IEEE Int. Conf. on Neural Network, vol. 3, pp. 1710-1715, 1996.