在本論文中，我們提出一在事件觸發式無線感測器網路(Wireless Sensor Network, WSN)下，基於速率分配及使用高可用性封包傳遞路徑之多路徑壅塞控制方式(multi-path congestion control approach)。本方式包含兩部分，第一部分為建立一以基地台(sink)為根節點，連接WSN中各感測器之生成樹，並以其為初始路由路徑。第二部分為藉由使用多路徑及速率分配，來建立一公平之封包傳遞環境。一般來說，新事件的發生會增加網路中的流量，將會使新建立的路由路徑上之節點產生雍塞，尤其是在基地台附近之節點，因而破壞原先流量之平衡並導致封包之丟棄。事實上，若干事件之訊息無法即時地傳遞到sink，可能會造成決策者因資訊不足而做出錯誤的決策。而在事件結束時，因停止傳送封包，亦將破壞路由上之中間節點的流量平衡。因此，我們動態地調整頻寬以維持各流量間之公平性和善加利用現有資源。實驗結果顯示本方法可有效的改善無線感測器網路的吞吐量、封包延遲及封包遺失率。 In this paper, we proposed a rate-allocation based multi-path congestion control approach, called multi-path-congestion control method (MUCON for short), which enforcing high path availability of packet delivery for event-driven Wireless Sensor Networks (WSNs) consists of two parts. The first is constructing a spanning tree to connect all sensor nodes of a WSN to the sink for initial routing. The second is establishing a fair packet forwarding environment by employing multi-path and rate control to deliver packets through routing paths with a balanced manner. Generally, the occurrence of an event will increase net flow, which may congest the nodes on the newly established routing path if the nodes are currently shared by several routing paths, particularly for those nodes near the sink, consequently ruining originally balanced traffic on paths and forcing some event packets to be dropped. In fact, without the information conveyed on the lost packets, users may make an inaccurate decision and react improperly for events. When an event disappears, the surrounding sensor(s) stops transmitting packets. This will again destroy the fairness and unbalance net flows flowing through the co-node routing paths. Thus, to maintain the fairness and balance net flows for a WSN, a downstream node needs to dynamically adjust bandwidths for all its upstream nodes. Experimental results show that this method can effectively improve a WSN’ throughputs, shorter end-to-end delays and reduce packet loss rates.
