目標物覆蓋(Target Coverage)在無線感測網路(Wireless Sensor Networks, WSNs)中是個很重要的研究議題。然而，在過去的研究中，大都假設目標物間相互連結。本論文所考慮的目標物(Target Points)乃分散於許多不連續區域中。為使目標物的資訊可在時限內傳達至收集站(Sink Node)，行動資料蒐集裝置(Data Mule)必須設法以移動性來達成網路的連通及目標物資訊的收集等目的。本論文提出了Time-Constrained Target Points Patrolling (TCTP) Algorithm，使分處於各地的Data Mule以分散式的方式，在目標物之間建置出一有效率的巡邏路徑，進行感測及資料蒐集的任務，並使這個路徑中每個目標物具有相近的資訊更新時間週期，並考慮Data Mule的電量限制與充電問題。此外，本論文亦針對具不同權重的目標物提出Weight-TCTP(W-TCTP) Algorithm，使Data Mule對於重要性較高的目標物，給予其較高的感測與收集頻率，提高其資料更新之速度。 除了收集目標物的路徑建外，本論文亦針對Data Mule如何在初始階段快速到達目標物進行深入研究。本論文考慮一行動感測網路之派遣問題，每個目標物依重要性不同的監控品質(Quality of Monitoring, QoM)需求，但由於行動感測器在一開始隨機散佈於環境中，使行動感測器之間並無法利用通訊能力協調所要前往的目標物，這使得行動感測器的派遣面臨極大的挑戰。本論文考慮目標物分散在各區域且不連續，為了對這些不連結的目標物進行監控，本論文發展一有效率的Mobile Sensor Dispatching Mechanism (MDM)，可使隨機散佈於環境中的行動感測器以最少的移動成本迅速的覆蓋目標物，且滿足目標物之QoM需求。此外，本論文還提出了Learning Dispatching Mechanism (L-MDM) ，使行動感測器在執行派遣工作的同時，利用與其他感測器短暫交會的機會，分享其移動經驗，使行動感測器透過可學習機制，以更高的效率覆蓋目標物且滿足目標物的QoM需求。 Target coverage problem has been widely discussed in Wireless Sensor Networks (WSNs). Different from the traditional target coverage problem, this paper considers the scenario that the target points are distributed over several disconnected areas. To collect the target information within a given time period, data mules should move and visit all target points for data collection in a disconnected network. Since the time interval (also referred to visiting interval) for consecutively visiting to each target point reflects the monitoring quality of this target point, the goal of this research is to minimize the maximal visiting interval for each target point. Given a number of data mules, this thesis initially proposes a Time-Constrained Target Points Patrolling (TCTP) algorithm which aims at constructing an efficient patrolling path for all data mules to visit each target point with a minimal visiting interval. Then a Weighted-TCTP (W-TCTP) algorithm is further proposed to cope with the weighted target problem where some targets might have higher weights and should have higher data update frequencies. In addition, this thesis consider the energy constraint of data mules and additionally proposes a RW-TCTP algorithm that treats energy recharge station as a weighted target and arranges all data mules visiting the recharge station before exhausting their energies. Simulation results reveal that the proposed TCTP and RW-TCTP mechanisms outperform the existing work in terms of visiting interval. In addition to the path construction problem, this thesis also considers the dispatching problem. In the considered environment, each target point gi is assigned with a Quality of Monitoring (QoM) value QoMi which representing the number of data mules required for monitoring the target. The given mobile sensors are randomly distributed in the monitoring region and thus they can not communicate with each other. As a result, the key challenge of the dispatching problem is that each mobile sensor should locally construct a visiting path passing thourhg all targets such that all targets can rapaidly satisfy their own QoMs with a minimal movement cost of mobile sensors. To cope with the dispatching problem, this thesis presents two efficient Mobile Sensor Dispatching Mechanisms (MDM) which aims to minimize the movement cost of mobile sensor while the QoM demand of each target is rapaidly satisfied. Experimental study shows that the proposed dispatching mechanism has a better performance than existing work in terms of movement cost of mobile sensors.