ZigBee無線感測器網路之網路議題研究與探討

Abstract

ZigBee為一套新興的通訊協定標準，其被制定為適合低功耗且低傳輸率之無線感測器網路。在ZigBee網路協定堆疊中，其實體層及媒介存取層協定皆採用IEEE 802.15.4之規範。ZigBee更於IEEE 802.15.4之上，規範了網路層至應用層，因此，ZigBee解決了不同無線感測器平台之間的互通性的問題。在ZigBee標準協定中，IEEE 802.15.4制定了兩種不同的媒介存取模式：信標模式(beacon-enable mode)及無信標模式(beaconless mode)；而ZigBee於網路層規範了樹狀網路(tree network)及網狀網路(mesh network)。本篇論文便主要針對於ZigBee之樹狀網路及網狀網路中，提出三項研究主題。

由於ZigBee信標模式下之樹狀網路提供了省電機制及輕量之路由協定，因此其被視為於大家所熟知的ZigBee網路拓樸中，特別適合無線感測器應用之網路拓樸。此拓樸之樹狀骨幹是透過ZigBee規範中之分散式位址分派演算法所形成，此位址分派演算法擁有簡單且容易實行的特性，但它必須於網路生成前，事先規範網路中路由節點之最大可容納之小孩(child)數及網路之最大深度(depth)。我們觀察到其ZigBee位址分派演算法規則太過嚴謹，因此造成位址空間使用率不佳，使得網路節點在位址空間足夠的情況下，仍因這些限制而無法連上網路而形成孤兒(orphan)節點，而ZigBee網路中之孤兒問題將會導致無法很容易地增加網路覆蓋率。因此，於本論文的第一個研究主題將探討如何有效率地降低ZigBee信標模式樹狀網路之孤兒節點。於本論文中，我們利用了時間性上位址重覆使用的概念，設計了一相容於ZigBee規範(ZigBee-compatible)之位址分派演算法來有效地降低孤兒節點數並提高網路生成量。同時，於此時間性上位址重覆之網路架構中，我們也設計了一個輕量級且以網路位置為基礎之樹狀路由。最後，利用模擬實驗的結果，也驗証了我們所設計之方法的確可以有效地降低網路孤兒數。

而於第二個研究主題方面，我們進一步地探討ZigBee位址分派演算於長鏈狀網路中所遭遇之孤兒問題。雖然於本論文第一項研究中，我們利用時間性上的位址重覆使用的概念，有效的解決了孤兒問題，然而由於ZigBee網路生成演算法之規範，使得網路生成仍然是一個很重要的議題。同時，我們觀察到在許多的無線感測器應用中，其網路拓撲經常形成一path-connected-cluster (PCC)拓撲。此拓撲中包含多數個需要較密集的感測器進行感測之叢集網路(cluster)，而這些叢集並無實體連結性，因此需要利用感測器佈建成長鏈狀來完成其連通性，而這種網路拓撲我們定義其為PCC-WSNs，而且許多的實際的感測器網路應用經常形成此PCC-WSN拓撲，如經橋樑連結之群島的感測或經由多個小徑連結之建築物監控等。而本論文將於此網路拓撲中，我們說明了ZigBee位址分派演算法於此PCC-WSNs中將會極差之表現，因此，本論文將提供一系統化之解決方案，包含了網路生成協定、自動位址分派演算法，及輕量地路由協定，同時，經由模擬結果，可以驗証我們所設計之方法於PCC-WSNs網路生成上是非常有效率的。

在最後一個研究主題裡，我們探討了ZigBee網狀網路形態中所存在的問題。網狀路由提供了一個非常有彈性且健全之多跳通訊路由協定。於ZigBee網狀網路規範中，其利用了隨機式位址分派演算法解決網路生成孤兒問題，然而ZigBee規範其網狀網路只能運行無信標模式之媒介存取協定，因此，省電將是ZigBee網狀網路中一個極為重要之議題。所以，本論文利用跨層(cross-layer)設計的概念，透過非同步式之省電媒介存取層協定，於ZigBee網狀網路中設定一有效的省電路由協定。所以於媒介存取層上，我們將使用grid-quorum架構，使每個節點有一個醒睡的機制；而於網路層中，我們將考量到一路由路徑上，每個節點於quorum機制下所消耗的電力，調整路徑節點之quorum形態，選擇出較為省電之路由路徑。最後，實驗模擬結果也將驗証我們所設計的方法的有效性及可行性。

ZigBee is a communication standard which is considered to be suitable for wireless sensor networks. In the ZigBee protocol stack, physical and MAC layer protocols are adopted from the IEEE 802.15.4 standard [16]. ZigBee solves interoperability issues from the physical layer to the application layer. IEEE 802.15.4 defines two different modes for medium access: beaconenabled mode and beaconless mode. ZigBee supports tree and mesh network topologies. In this dissertation, we will focus on the networking issues in ZigBee beacon-enabled, tree-based networks and beaconless mesh networks.

Among the well-known ZigBee topologies, ZigBee beacon-enable cluster-tree is especially suitable for wireless sensor applications with its supporting of power-saving operation and lightweight routing. The backbone of a tree network is formed by ZigBee distributed address assignment scheme. This assignment is easy to implement, but it restricts the number of children of a device and the depth of the network. We observe that the ZigBee address assignment policy is too conservative, thus usually making the utilization of the address pool poor. Those devices that can note network addresses will be isolated from the network and become orphan nodes. The orphan problem leads to the difficulty in smoothly increasing the network coverage or device density. Therefore, we propose our first research work to addressing how to alleviate the orphan problems effectively. We propose a ZigBee-compatible address assignment through temporal duplications to alleviate orphans and scale the networks. A light-weight, address-based, tree-based routing is proposed to support one-to-one routings in address-reused environments. Simulation results verify the effectiveness of the proposed solution.

In our second research work, we discuss the ZigBee orphan problems in long-thin (LT) topologies. Although our temporal duplication addressing can significantly alleviate orphans in ZigBee tree networks, its deployment is still a main concern. We further observe that many monitoring applications for WSNs have adopted a path-connected-cluster (PCC) topology, where regions to be monitored are deployed with clusters of sensor nodes. Since these clusters might be physically separated, paths of sensor nodes are used to connect them together. We call such networks PCC-WSNs. PCC-WSNs may be widely applied in real situations, such as bridgeconnected islands, street-connected buildings, and pipe-connected ponds. In this work, we show that the address assignment scheme defined by ZigBee will perform poorly in terms of address utilization. We then propose a systematical solution, which includes network formation, automatic address assignment, and light-weight routing. Simulation results verify the effectiveness of the proposed solution.

In our third research topic, we address the networking issues in ZigBee mesh networks. Mesh network topologies are seen as a flexible and robust manner to provide multi-hop communication. Mesh topologies offer flexibility and robustness by facilitating path formation from any source to any destination within the network. In ZigBee, stochastic address assignment mechanism is recommended in ZigBee mesh topologies such that networks can be easily scaled up without orphan problems. However, ZigBee mesh networks can only operate in beaconless mode. Power saving is a major concern in a wireless sensor network. Hence, we are interested in linking the asynchronous power-saving protocol and the energy-efficient routing problem together in ZigBee mesh networks. A cross-layer solution is proposed. On the MAC layer, we propose to use the grid-quorum system [40] to serve as the underlying power-saving framework. On the network layer, we propose to find routing paths based on the power cost incurred by grid quorums used by nodes along a path. We show how these two layers interwork with each other to support continuous queries in an energy-efficient way. Simulation results also verify the effectiveness of the proposed solution.