颱風時雨量行動裝置預測系統之研發

本研究結合聚類分析(Cluster Analysis)理論與調適性網路模糊推論系統(Adaptive Network-Based Fuzzy Inference System, ANFIS)，建立颱風降雨預報模式(Typhoon Rainfall Forecasts，簡稱TRF)。並結合智慧型型行動裝置(Application，簡稱App)進行線上即時操作模擬即時之颱風降雨。研究中首先將歷史颱風事件透過聚類分析結果建立TRF模式最佳參數。以Qt Framework開發TRF Service與 TRF行動裝置App，TRF Service為TRF之網路服務，利用TRF App的UI介面進行操作，以Socket連線至TRF Service進行即時颱風降雨預測模擬。以淡水測站做為系統實際案例，經模擬分析顯示，將颱風事件以聚類分析進行分類擁有較佳的模擬結果，並且透過智慧型行動裝置的結合，增加即時模擬颱風降雨的便利性。亦提供歷史模擬記錄之查詢，可對不同模擬結果加以分析，不需再次重新模擬。本研究所使用之Qt Framework為一個以C++語言進行開發的Open Source軟體，具有跨平台特性，可由Qt Transformation system轉換成不同平台之智慧型行動裝置App。

關鍵字

調適性網路模糊推論系統；聚類分析；智慧型行動裝置；颱風降雨預測

並列摘要

This research integrates Cluster Analysis (CA) and Adaptive Network-Based Fuzzy Inference System (ANFIS) to build the typhoon rainfall forecasting model (TRF). It is incorporated into a mobile device application (App) to simulate online real-time typhoon rainfall. The historical typhoon events are used by CA to establish the optimized parameters of TRF. A TRF Service, a web service providing typhoon rainfall prediction, and TRF App are developed based on the Qt Framework. Real-time typhoon rainfall can be obtained by TRF App connecting to TRF Service through socket connections. Tamsui station is selected as the actual case to demonstrate the proposed approach and developed TRF Service and APP. The Simulated results show that applying CA to historical rainfall events to optimize the parameters of TRF can obtain more accurate typhoon rainfall. It is more suitable for typhoon rainfall prediction by combining TRF with App technology; it is also convenient that users can analyze and compare many historical simulated results provided by TRF App without re-simulating them. In this study, the Qt transformation system, a tool of the Qt Framework (a well-known cross-platform, open source C++ UI library), is applied for generating native App on different platforms easily.

並列關鍵字

Adaptive Network-Based Fuzzy Inference System ； Cluster Analysis ； smart mobile devices ； Typhoon Rainfall Forecasting

參考文獻

14. 沈宏榆，2013，” 整合聚類分析與類神經網路於區域淹水預測之研究”，淡江大學研究所碩士論文

16. 張斐章，張秋麗，2006，類神經網路

18. 陳泓碩，2012，” 應用ARIMAX及ANFIS模型於福山森林集水區逕流模擬之研究”，台灣大學森林環境暨資源學研究所碩士論文

1. Chu, H.J., Liau, C.J., Lin, C.H., and Su,B.S.,2012,” Integration of fuzzy cluster analysis and kernel density estimation for tracking typhoon trajectories in the Taiwan region’, Expert Systems with Applications, Vol.39, Issue 10, pp.9451-9457

2. Chiu, S., 1994,”Fuzzy model identification based on cluster estimation”,J. Intell. Fuzzy Syst., Vol.2 , pp.267-278

國際替代計量

颱風時雨量行動裝置預測系統之研發

未授權

主題瀏覽