在無線感測網路中以Quorum為基礎處理多個查詢的省電繞徑演算法

Abstract

隨著無線感測網路相關的研究議題得到許多研究單位及學者的關注，近年來許多適用於無線感測網路，以省電為目標的MAC層與網路路由層通訊協定相繼發表，但現有的研究成果，絕大多數皆假設感測網路的節點具備有時間同步的條件，才能顯示其省電的效果；然而，節點們彼此要具備時間同步本身又是另一個相當耗費能源的研究課題，特別是在無線感測網路這樣的多跳通訊環境，尤其困難。之前的研究已經證明了，使用Quorum系統的特性可以在不具有時間同步的無線行動通訊環境下達成省電的設計；因此在這篇論文中，我們提出一個以Quorum系統為基礎的通訊協定，在無線感測網路中處理多個持續性查詢的同時，計算出每個節點不同的醒睡排程以節省能源不必要的浪費。另外，我們也設計遶徑演算法，使節點能夠動態調整Quorum系統來改變本身的醒睡排程，儘可能用較低的能源消耗去配合持續性查詢的流量需求。本篇的貢獻可以分為三個方面：首先，我們提出一個涵蓋MAC層與網路路由層的跨層系統設計，不同於以往只針對其中一層設計的研究，本篇論文利用Quorum系統適切地結合MAC層與網路路由層，在無線感測網路中完成支援多個持續性的查詢。第二，我們說明了如何利用簡單且不需龐大代價的局部同步演算法，協調在同一路由路徑上多個節點的Quorum系統來達到能源保存。第三，由於一個感測節點可能遇到必須滿足不同持續性查詢的不同流量需求又要保持較低的能源消耗的情況，我們提出讓節點能夠適應性地支援多個Quorum系統的方法。模擬的結果也指出我們的方法的確可以得到較低的能源消耗，並且有效地完成多個持續性查詢的工作。

Recently, several energy-efficient MAC and routing protocols have been proposed for wireless sensor networks (WSNs). Most existing schemes require nodes to be synchronized in time. However, synchronization is costly, especially in multi-hop communication environments. Quorum systems have been shown to be able to support power-saving design in a wireless and mobile environment without requiring time synchronization. In this work, we propose a quorum-based protocol to derive their wake-up schedules to support continuous queries. We also propose a routing scheme to help nodes dynamically change their quorum patterns to meet continuous queries' traffic demands at low energy costs. The contributions of this work is three-fold. First, our design is a cross-layer design covering MAC and routing layers. While most existing protocols only address one of these layers, our work nicely tailor MAC layer using quorum systems with routing layer to support continuous queries. Second, we show how to coordinate quorums of multiple nodes in a routing path for energy conservation. A simple, inexpensive, local synchronization among nodes is proposed. Third, we show how to adaptively support multiple quorums in a sensor node when it is passed by several query routing paths; this can effectively meet continuous queries' traffic demands while keep energy consumption low. Simulation results indicate that our scheme can effectively support continuous queries at low energy cost.