由於生態監測系統需要進行長期監測及擷取連續生態環境資料,才可以透過所收集的資料分析並建立生態模型。因此,本論文依照生態監測之特性提出一改良型的無線感測器網路MAC協定,達到長時間的監測及穩定的資料收集。 本協定以sensor-MAC (S-MAC)的週期性工作及睡眠週期為基礎概念,透過節點間同步對時機制使節點同步地進入工作及睡眠週期。在節點的資料傳輸流程中,考慮節點層數及時槽分配兩種因素分配節點的工作排程,減少資料碰撞及工作週期的閒置時間。此外,本協定特別提出方法解決感測器網路在資料收集時發生遺失的問題;由節點監視資料傳輸的狀況,當發現有遺失或未傳送之資料,另外以多重路徑傳輸的方式補傳資料。當節點因為電源耗盡或其他原因產生故障時,節點透過收集鄰近節點的資訊,自行判斷一有效通訊之傳輸路徑,保持無線感測器網路連線的完整。 最後,本論文分析並測試本協定之電源消耗、資料穩定性及傳輸延遲等性能指標,其性能比S-MAC佳並較具適應性。除此之外,本論文將協定實地應用於東方果實蠅生態監測系統上,證實本協定之實用性並擁有較好的效能。
The ecological monitoring system usually needs to monitor the targets for a long time, and acquires measured data steadily. Only contented with the conditions, the data just can be used to establish the ecological model. Therefore, this thesis presented an improved medium-access control protocol for wireless sensor network that aims to content with the characteristics of ecological monitoring, and it is applicable for a long-term monitoring and collecting reliable data. The proposed protocol adopted the concept of active/sleep mode in Sensor MAC (S-MAC). The sensor nodes used synchronization protocol to make nodes active/sleep synchronously. The protocol considered the number of network hierarchical levels and time slot assignment to plan the node working schedule during data transmission. It can improve the problem of data collision and idle nodes. Besides, the protocol monitored the situation of data loss while the wireless sensor network is transmitting data. When there were lost or un-transmitted data, the nodes uploaded the data by multi-paths to reduce the data loss and complete data collection. In order to keep the reliable connection of the wireless sensor network, the protocol demanded nodes to collect the information of the neighbors and found an alternative path when some nodes were out of power or damaged. Finally, we compared the designed protocol with S-MAC protocol in the power consumption, data stability and transmission latency. We also used the designed protocol in application of the oriental fruit fly monitoring system to examine the efficiency and functions.