近年來，由於無線多媒體感測網路中多種應用的興起，使得其成了熱烈討論的議題，而如何在充滿限制的環境下，提升其效能與生命週期已成了重要的焦點。 相較於單一最佳化路徑，多重路徑更適合用在無線多媒體傳輸環境下，因為它可以增加頻道使用率、降低傳輸延遲、還可以平均能源的負擔，讓無線多媒體傳輸能更有效率。但是在建立多重路徑時，我們必須同時將無線網路的環境和感測節點的特性列入考量，否則將無法建立出合乎效益的多重路徑。 在本論文中，我們會先解釋無線多媒體感測網路中，資料傳輸架構之特性，以及使用多重路徑的利與弊，尤其是多重路徑時，路徑間干擾的問題。接著探討之前文獻中所提出的路徑演算法們的優缺點，和目前所遭遇到之問題，然後統合以上幾點，提出一針對拓樸快速變動、且能源有限之無線多媒體感測網路下的不干擾多重路徑演算法。 在我們提出之不干擾多重路徑演算法中，只需要將整個網路分成數個區域。並在同時間內使用不同的區域傳送資料，就可以輕易達成不干擾之多重路徑傳輸。 每次傳送資料時時，封包中也會附帶轉傳節點的能源資訊及傳輸時的邊界資訊。讓目的端節點將這些資訊回報給來源節點，使來源點可以根據各個區域的能源狀況及節點密度，調整各區域的範圍及使用率，藉此讓我們的多重路徑、能因應感測網路中節點的能源狀況和拓樸變動等調整其傳輸路徑，使得我們的機制能在嚴苛的感測網路環境中，有效率地傳輸多媒體資料。 實驗證明，我們的機制所建立出來的多重路徑，能夠比其他機制適合傳輸即時多媒體。且能平均的將傳輸資料的重擔分散至整個網路，延長電力有限的無線感測網路之生命期。 Wireless multimedia sensor networks have been widely discussed in recent years because there are many potential applications, and how to increase the throughput in constrained environments has become the main challenge. Multipath routing is a promising solution because it can improve the channel utilization rate, reduce transmission delay, and balance the transmission load, but the efficiency will be limited if we construct the paths without taking the characteristics of wireless environments and sensor networks into consideration. In this thesis, we first explain the characteristics of data transmission schemes in wireless multimedia sensor networks, and then analyze the pros and cons of multipath routing, especially the problem of inter-path interference. After discussing the related works about routing schemes, we propose a Geographic Energy-Aware non-interfering Multipath (GEAM) routing scheme which can keep high performance through harsh sensor network environment. The proposed scheme divides the whole network topology into many districts, and in the same time forwards data through districts that won’t interfere with others, to achieve interference-free transmissions. We also adjust the load and boundary of each district according to the energy and location information of nodes in that district. Our scheme does not construct fixed paths in advance and can keep high performance even though the network topology changes rapidly. In the simulation we show that proposed scheme achieves high performance for real-time transmission and can evenly distribute the transmission loads to most of nodes in the network.
