可充電無線感測網路(Wireless Rechargeable Sensor Networks, WRSNs)可利用無線充電器(wireless charger)適時補充裝有無線充電接收器(harvester)之感測節點的電力，以維持所有感測節點持續運作，使其具有永續性(sustainability)。本論文探討可充電無線感測網路中無線充電器的佈建與排程。在佈建無線充電器的目標為使用最少無線充電器佈建於無線感測網路環境之中，並保證所有感測節點的永續性；排程無線充電器的目標為設定個別充電器的啟動比例(active ratio)，使得所有感測節點的剩餘電量穩定並維持系統的高充電效益，減少充電器不必要的電力浪費。本論文針對以上兩個議題分別提出： (1) 適應點對貪婪圓錐選取演算法(Adaptive Pair-Based Greedy Cone Selection, APB-GCS)，以每對節點產生的圓錐和該充電器對各感測節點的充電效益作為考量基礎，用貪婪的方式選擇最少數量的充電器覆蓋環境中的所有感測節點，並計算充電效益是否滿足所有感測節點的充電需求，使WRSN達到永續性。(2) 高充電效益優先排程演算法(High Charging Efficiency First Scheduling, HCEFS)，可根據充電效益計算出各充電器工作比例，以貪婪方式優先挑選每個充電器有效充電範圍中感測節點總耗電最大的時刻進行工作排程。 我們使用Powercast P2110-EVAL-02充電器設備進行實驗，用以求得充電器對感測節點在不同距離、水平角度、垂直仰角的電功率。透過乘冪迴歸分析(power regression analysis)得到與福利司模型(Friis model)相似的電功率乘冪回歸方程式，作為逼近真實充電效益的數據並進行模擬。模擬結果證實APB-GCS相較於相關方法可以使用較少數量的無線充電器佈建且滿足所有感測節點的需求。此外，我們也模擬了充電器的工作比例及系統總耗電量，結果證實HCEFS在確保WRSN永續性的條件下可達成節省充電器總耗電量的目標。 ;Wireless Rechargeable Sensor Networks (WRSNs) use wireless charging technologies to supply sensor nodes’ power and maintain sustainability. This thesis discussed the deployment and scheduling of the wireless chargers in WRSNs. The target of wireless charger deployment is to use the fewest chargers to guarantee the sustainability of all sensor nodes. In addition, the target of scheduling is to calculate and schedule chargers’ active ratios, resulting in the stability of all sensors’ residual power, maintaining high charging efficiency in the system, and reducing the unnecessary power resources abuse. To solve the issues above, we proposed: (1) Adaptive Pair Based Greedy Cone Selection (APB-GCS), using adaptively adjusted node pairs to calculate the charging efficiency, to minimize the number of wireless chargers and fulfill all the sensors’ demands, and (2) High Charging Efficiency First Scheduling (HCEFS), based on the charging efficiency, to calculate the active ratio of every charger, and greedily select the timeslots of the most power consumption for scheduling chargers. We used Powercast P2110-EVAL-02 device to experiment the charging efficiency of every pair of chargers and sensors between different distances, different horizontal and vertical angles, and modeled the experiment results through power regression analysis to get power regression equations which are similar to the Friis model. We then took the equations in our simulation to derive approximately the charging efficiency. The simulation results show that we can use fewer chargers to fulfill all sensors’ energy demands. We further performed simulation to evaluate the chargers’ active ratios. The simulation results show that we can achieve the target of reducing chargers’ power consumption in sustainable WRSNs.
