智慧型保修系統資料擷取抽樣與傳輸計劃最佳化之模型研究

The Modeling of Data Acquisition Sampling and Transmission Planning for Intelligent Maintenance System

李英凱(Ying-Kai Lee)

指導教授 ： 邱昭彰

元智大學/資訊學院/資訊管理學系/碩士(2003年)

摘要

國軍之新一代武器日趨精密複雜，而武器維修人員卻因為役期的縮短而使得武器維修的動作因此而受到影響；在資訊科技與人工智慧均有長足進步的現代中，開發一個具有自動偵測武器設備與故障預警的資訊系統確有其必要性。智慧型保修系統是一個超高壓電源裝置的監測系統，提供使用者在遠端觀測受測裝置中的重要線上數據。這些數據經由資料自動學習機制的分析，可找出系統故障前這些數據與系統狀態間的關係模式（pattern）。此模式可預先推測系統組態，並於確認警訊後立即透過網路傳輸通知維修人員檢測裝置，以提高裝備的可靠性及完善率。然而欲達此目的，其系統成敗關鍵仍繫於感應器資料蒐集取樣與網路傳輸頻寬的有效規劃。
但由於傳統方法對於決定資料刪減策略與分配網路頻寬權重這兩方面並不能提出有效之解決方案，因此，我們計劃將針對整個保修系統進行下列部分改進之道，期以透過最佳化的規劃，並達到系統建置成本降低及整體效率提升的目的：
1. 資料傳輸之網路頻寬比例分配最佳化。
2. 資料刪減策略之最佳化。
我們期以透過最佳化的規劃，達到建置成本降低及整體效率提升。

關鍵字

智慧型保修系統 ； 資料擷取 ； 基因演算法 ； 蒙地卡羅模擬 ； 模擬最佳化

並列摘要

The modern military weapons are getting much more complicated today. The military service periods of maintenance personnel, however, are getting shorter and shorter. Therefore, the activities of weapon maintenance are deeply affected. This is a very era when information technology and artificial intelligence both get evolved, and it is necessary to develop information system [32] that can automatically detects weapon devices problems and alarms before problems occur. Intelligent maintenance system is a monitoring system and provides users with important values of remote detected devices. Those values will be analyzed by data automatic learning mechanism to find out patterns between values and the device status before the device fails. The pattern can inform maintenance personnel immediately via network after confirming alarms to enhance devices reliability. To achieve this goal, however, the key point is relied on the effective planning of data acquisition sampling and transmission.
Due to the traditional algorithm’s failure in determining both the discarding strategy and network bandwidth allocation within the system, the proposed research propose two methods to improve the known difficulty. The two methods are as follows:
1. The optimization of data transmission network bandwidth allocation, and
2. The optimization of the server queue discarding strategy
We expect to use the optimization planning technique to reduce the system development costs and to enhance the overall efficiency.

並列關鍵字

Intelligent Maintenance System ； Data Acquisition ； Genetic Algorithm ； Monte Carlo Simulation ； Simulation Optimization

參考文獻

[2] A. Ramesh, A. Chockalingam, L. B. Milstein, “SNR estimation in Nakagami fading with diversity for turbo decoding,” Military Communications Conference, 2001. MILCOM 2001. Communications for Network-Centric Operations: Creating the Information Force. IEEE , Volume: 2 , 2001, Page(s): 1141 -1145 vol.2

[3] A. Ramesh, A. Chockaligam, L. B. Milstein, L.B, “SNR estimation in generalized fading channels and its application to turbo decoding,” Communications, 2001. ICC 2001. IEEE International Conference on , Volume: 4 , 2001, Page(s): 1094 -1098 vol.4

[4] A. Zeira, P. M. Schultheiss, “Realizable lower bounds for time delay estimation. 2. Threshold phenomena,” Signal Processing, IEEE Transactions on , Volume: 42 Issue: 5 , May 1994, Page(s): 1001 -1007

[5] B. Sadoun, “Applied System Simulation: a Review Study,” Information Sciences, Vol. 124, Issue. 1-4, pp. 173-192, May 2000.

[7] D. B. Fogel, “A Comparison of Evolutionary Programming and Genetic Algorithms on Selected Constrained Optimization Problems,” Simulation, Vol.64, Issue.6, pp. 399-406, 1995.