適用於農業之無線行動感測網路技術之研究---子計畫二：無線感測網路資料處理機制於農業應用

Abstract

隨著無線感測網路技術的進步，未來生活環境中，將佈滿感測器，利用無 線網路互相連結，並且負責監測並蒐集和生活環境相關的資料，此一環境架構為 無線感測網路環境。因此，利用無線感測網路於農業之應用，如農場的環境監控、 植物生長測量等，將是無線感測網路極為重要應用。然無線感測網路具有低耗電、 不具龐大記憶單元、低無線網路通訊頻寬及硬體運算能力較弱等限制。因無線感 測器之限制，使得在無線感測網路上的資料處理、查詢與傳統分散式系統有所不 同，且這些限制使得如何在有限的資源內，提供低耗電，有效率之資料處理與查 詢成為日漸重要的課題。 本計畫為一為期三年之總計畫『適用於農業之無線行動感測網路技術之研 究』之一子計畫。在此子計畫中，我們將研發於無線感測網路系統中，確保資料 正確，節省電源之無線感測網路資料處理機制。在本子計畫中，主要的研究課題 包含：(1)提供確保感測資料精確性機制；(2)有效率省電資料查詢處理;(3)資料收 集與探勘架構。首先，有鑑於感測器可能會有損壞，導致資料的不正確，我們將 研發過濾錯誤資料機制，以確保感測資料的正確性。其次，我們將研究多個資料 查詢最佳化及近似資料查詢機制，以降低感測器的電源耗損。最後，我們將建構 一階層式資料收集與探勘架構，以便有效率的收集資料並資料探勘，將有助於瞭 解感測資料所隱含之意義。 近年來，無論工業界或是學術界對於無線感測網路均相當的重視並且投入 大量的人力及研究，無線感測網路於農業應用極為適合，我們相信此計畫之執行， 將可研發出適用於無線感測網路系統中，資料處理前瞻性技術，並實作於農場環 境以作為學理與實務驗證。

The rapid progress of wireless communication and embedded micro-sensing MEMS technologies has made wireless sensor networks possible. It is expected that wireless sensors will be widely deployed for measurement, detection and surveillance applications. Due to the very nature of wireless sensor networks, applying wireless sensor networks for agricultural fields is deemed an important application. Note that sensors have several resource constraints such as power consumption, limited computing capability, smaller storage space and low bandwidth in communications. These resource constraints on sensors call for the design of data management in wireless sensor networks. This project is under the integrated project 『Wireless Mobile Sensor Network Technologies for Agricultural Applications 』, and aims at designing reliable data management and mining mechanisms in wireless sensor netowrks. Our primary goals include (1) developing data accuracy guaranteed mechanism, (2) devising energy efficient data query processing methods, and (3) exploring the feature of heterogeneous sensor nodes for data collection and in-network mining. More specifically, we will first develop trust relations among sensor nodes and explore trust relations for filtering faulty sensor nodes so as to ensure the correctness of data. Next, by exploiting the similarities in both data queries and sensing data, we will develop two novel query processing mechanisms for power saving. Finally, in order to discover the knowledge of sensing data, we shall develop a hierarchical data collection and mining architecture. In light of the hierarchical architecture, we will develop in-network mining algorithms. Through the collaboration with other sub-projects, a prototype of wireless sensor networks for agricultural applications will be built. In view of the increasing attention on wireless sensor networks, we strongly believe this project is very timely and will deliver results of both theoretical and practical importance.