標題: 行動無線感測網路之高效率定位技術
Efficient Localization in Mobile Wireless Sensor Networks
作者: 謝宜玲
Yi-Ling Hsieh
王國禎
Kuochen Wang
資訊科學與工程研究所
關鍵字: 動態參考;定位技術;行動無線感測網路;dynamic reference;localization;mobile wireless sensor network
公開日期: 2004
摘要: 在無線感測網路中,許多應用都需要定位技術。大部分現有的定位方法都是針對固定的無線感測網路。但在行動環境中,感測節點的位置就需要被定期更新。在本篇論文中,我們提出一個高效率定位技術,名為「動態參考定位技術」(DRL)。DRL是一個分散式的定位方法。為了節省定位時所耗的通訊成本,DRL藉由動態地改變參考點的氾濫段數(flooding-hop),來將參考點氾濫限制於局部區域,以降低節點之間資訊的氾濫。如此也使得行動節點只使用附近的參考點所提供的資訊,而非使用網路中所有的參考點,因此可提高定位正確性。此外,DRL是range-free的定位方法;因此它不需要特殊的硬體支援,例如訊號強度測量器、超音波量距器、或是具方向性的天線。DRL允許所有的節點都可以自由移動,其中只有少數的節點具有自我定位的能力(稱之為參考點)。DRL也能夠適應於低或高的節點密度,它採用動態參考點氾濫以及強韌的三角定位法。動態參考點氾濫會根據參考點周圍的節點密度來改變它的氾濫涵蓋範圍。而強韌的三角定位法允許參考節點不足的情況。我們已經評估了DRL與MCL。模擬結果顯示DRL的定位準確度比MCL[8]高出26%,尤其是在低密度參考點的情況下更顯得優異。藉由行動定位,DRL適用於電子地圖導航系統以及社區健康照護系統。
Localization is broadly required in many kinds of applications in wireless sensor networks (WSNs). Most existing localization methods are targeted at fixed WSNs. In mobile environments, the location of each sensor node needs to be updated at every certain interval. In this thesis, we propose an efficient localization approach, called Dynamic Reference Localization (DRL). DRL is a distributed localization approach. In order to save communication cost in localization, DRL reduces information flooding among nodes by dynamically changing each seed’s flooding-hop to limit seed flooding in a local area. In this way, mobile nodes use only information propagated from surrounding seeds, instead of using all seeds in the network; therefore the location accuracy can be improved. In addition, DRL is a range-free approach; thus it does not require any special hardware supports, such as signal strength measurement, ultra-sound ranging, or directional antennas. And DRL allows all the nodes mobile and moving freely, while there is only a limited fraction of nodes having self-positioning capability (called seeds). Moreover, DRL can adapt to low or high node density, because of its dynamic seed flooding and robust triangulation. Dynamic seed flooding is to change flooding coverage according to each seed’s surrounding node density. Robust triangulation allows the cases of insufficiency of reference nodes. We have evaluated DRL and MCL[8]. Simulation results have shown that the location accuracy of DRL is 26% higher than that of MCL. Especially in low seed density condition, DRL outperforms MCL even more. With mobile positioning, DRL is suitable for applications, such as navigation systems using e-map and community health-care systems, in outdoor environments.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT009223505
http://hdl.handle.net/11536/76553
Appears in Collections:Thesis


Files in This Item:

  1. 350501.pdf