無線感測網路(Wireless Sensor Networks)已經被視為一項未來的偵測技術，將大量的感測器節點隨機分佈後可以提供許多廣泛的應用，例如：追蹤某種目標物之狀態、環境之監測、週遭安全性之提升等。在無線感測網路中，因為感測器節點只能使用有限的能源的關係，例如電池，所以當感測器節點的能源耗盡時，會導致整個無線感測網路的功能受到影響，甚至造成整個網路無法運作。因此管理感測器的能源成為一個重要的議題，藉由有效的管理感測器的能源才能延長無線感測網路的生命週期(Lifetime)。 要節省能源最好的方式就是將感測器節點的電源直接關閉，但是被關閉的感應器節點就不再屬於網路的一部份，無法執行資料的傳輸與接收，可能會造成資料的遺失，因此何時該關閉電源何時該打開電源的權衡便顯得十分重要，一方面要延長無線感測網路的生命週期(Lifetime)，而另ㄧ方面又不能造成太高的資料遺失率，在“應用於無線感測器網路中以休眠排成為基礎之電源管控機制＂中所提出的休眠排程演算法(Sleep Scheduling)，因為其無法保證感測器節點是否有鄰近的感測器節點(neighbor node)可以代傳資料，明顯地容易造成資料的遺失。因此本篇論文將針對這個問題進行改善，提出以傳輸功率控制為基礎的改善休眠排程來解決資料遺失的問題。 依據論文中所提出的改善演算法與原有的演算法進行模擬實驗分析，發現原有的演算法可以完成43個Rounds，而且會有些許的資料遺失率，而改善的演算法可以完成40個Rounds優於沒有使用休眠排程演算法的35個Rounds，但沒有資料遺失的狀況。所以本論文所提出的改善演算法可以有效的解決資料遺失的問題。 Wireless sensor networks (WSNs) have become a sensing technology of future. A large number of sensor nodes can be randomly sprayed in an area and support many applications such as tracking the status for a target, monitoring environment and improving public security. The sensor nodes of WSNs can only use limited energy like battery. When the energy has been exhausted, WSNs cannot work anymore. In order to extend the lifetime of WSNs, the management of the sensor''s energy is an important issue. The best way to save energy is to shutdown the power of sensor node directly. The sensor node which the power is shutdown is not belonging to the WSN and cannot do the data transmission and receive anymore. To shutdown the power of sensor node may cause the data loss. So when to turn off the power and when to turn on the power is become vary important. Shutdown the sensor nodes may extend the lifetime of WSNs, but also may cause the high packets loss rate. The Sleep Schedule addressed on “A Power Control Mechanism based on Sleep Scheduling for Wireless Sensor Networks” cannot guarantee that neighbor node has ability to do transmission, because the node may be in sleep. And it may result in data loss. This thesis would use the transmission power control technology to improve this issue and reduce the packets loss rate. In this thesis, use the improved algorithm and the original algorithm to do simulation. The original algorithm can complete 43 rounds and has some data loss. The network without sleep scheduling algorithm can complete 35 rounds and has no data loss. The improved algorithm which is better than the network without sleep scheduling algorithm can complete 40 rounds and has no data loss. As a result, the improved algorithm proposed in this thesis can effectively resolve the data loss problem of the original algorithm.
