無線感測網路中的感測器多由電池供應電量，由於電池更換不易，容易導致感測器電量一旦消耗殆盡，感測器節點將無法運作，只能視同死亡，整體網路覆蓋面積範圍將逐漸縮小，使得監控品質很差，而死亡節點無法接收及傳遞資料，也將導致網路斷路，無法將監控的資料傳回，因此，如何維持感測器的電量，使其運作正常，一直是學者們所共同努力的目標。
本研究考慮在一個已佈署多個感測器的無線感測網路中，散佈多台行動充電車，行動充電車彼此之間可以互相通信。由於每一個感測器有不同的耗電率，以及不同的覆蓋面積，行動充電車將考慮感測器的覆蓋貢獻與距離去判斷如何移動，由於在網路當中散步多台行動充電車，因此感測器需要判斷應該向哪一台行動充電車發送充電請求，可以使其本身等待時間最短。本論文著重於最大化無線感測網路中所有感測器累積的覆蓋效益，探討感測器在有限的電量下，透過增強式學習的應用，規劃行動充電車充電路徑，以達到感測器覆蓋目標最大化。我們將增強式學習分為兩部分，第一，透過行動充電車的學習，可以學習如何移動將會獲得最大效益；第二，透過感測器的學習，得知向哪一台行動充電車發送請求將獲的最大效益。
本論文之貢獻為提升整體無線網路的感測器覆蓋面積，使行動充電車在判斷充電對象時，以獨立覆蓋面積較大的感測器優先進行充電，此舉可使整體的覆蓋面積維持大範圍的覆蓋，使整個網路不會因為某些區域的感測器死亡無法傳輸以及接收資料，使用增強式學習讓行動充電車與感測器分別學習應該如何規劃自身排程。

Most of the sensors in the Wireless Sensor Network are powered by batteries. The battery is hard to replace cause the sensor lose the power, and that will result in sensor node cannot work. It only can be regarded as death. The overall Wireless Sensor Network coverage will gradually shrink, making the monitoring quality very poor, and the dead node will not be able to receive and transmit data, and also cause the network to be disconnected and unable to return the monitored data. Therefore, how to maintain the power of the sensor to make it work normally, has always been the goal of the joint efforts of scholars, bile charging car to send the maximum benefit that the request will receive. The research considers that in a Wireless Sensor Network with multiple sensors deployed, multiple mobile chargers are distributed so that the mobile charger can communicate with each other. Since each sensor has a different power consumption rate and a different coverage area, the mobile charger will consider the coverage contribution and distance of the sensor to determine how to move. Because there are multiple mobile charger walking in the network, so the sensor needs to determine which mobile charger should send a charging request to, so that it can minimize its waiting time. This paper focuses on maximizing the coverage benefits accumulated by all sensors in the wireless sensor network, and explores how the sensors can be used with limited power and through the application of enhanced learning to plan the charging path of mobile charging vehicles to achieve the sensor maximize coverage goals. We divide the enhanced learning into two parts. First, through the learning of mobile charging cars, you can learn how to move to get the most benefit; second, through the learning of sensors, know which mobile charging car to send the maximum benefit that the request will receive. The contribution of this paper is to increase the sensor coverage area of the overall wireless network, so that when determining the charging object, the mobile charging car will be charged with the sensor with a larger independent coverage area first, which can maintain the overall coverage area Large-scale coverage prevents the entire network from being unable to transmit and receive data due to the death of sensors in certain areas. Using enhanced learning allows mobile charging vehicles and sensors to learn how to plan their own schedules.

目錄	VI
圖目錄	VIII
表目錄	IX
第一章、簡介	1
第二章、相關研究	5
第三章、網路環境與問題限制	8
3.1網路劃分限制	10
3.2行動充電車限制	10
3.3感測器限制	11
第四章、行動充電車增強式學習規劃	16
4.1 行動充電車增強式規劃	17
4.1.1分散移動	18
4.1.2隨機移動	22
4.1.3目標移動	24
4.2 感測器增強式學習規劃	31
第五章、模擬實驗	38
第六章、結論	42
參考文獻	43
附錄-英文論文	45

圖目錄
圖1行動充電車移動方向與感測器覆蓋範圍 11
圖2行動充電車之間互斥力 19
圖3邊界對於行動充電車的排斥力 20
圖4第一階段合力 21
圖5網路格子分數 22
圖6第二輪加分機制 23
圖7行動充電車移動分向 24
圖8感測器所在得網路格子對行動充電車發出充電請求 25
圖9行動充電車移動方向 26
圖10行動充電車移動方向之投影 26
圖11感測器及行動充電車發送座 標傳輸圖 33
圖12行動充電車與感測器狀態圖 37
圖13感測器場景假設圖 39
圖14感測器覆蓋面積比較 40
圖15充電延遲 40
圖16節點故障率 41

表目錄
表1相關研究比較圖 7
表2網路劃分相關定義 8
表3行動充電車相關符號定義 8
表4感測器相關符號定義 8
表5第四章符號相關定義 16
表6行動充電車狀態 17
表7行動充電車動作 18
表8行動充電車演算法 29
表9感測器狀態 32
表10感測器動作 32
表11感測器演算法 35

[1].	A. Mir and A. Khachane, "Sensing Harmful Gases in Industries Using IOT and WSN," 2018 Fourth International Conference on Computing Communication Control and Automation (ICCUBEA), Pune, India, 2018, pp. 1-3, doi: 10.1109/ICCUBEA.2018.8697380.
[2].	S. Zhang, J. Wu and S. Lu, "Collaborative mobile charging for sensor networks," 2012 IEEE 9th International Conference on Mobile Ad-Hoc and Sensor Systems (MASS 2012), Las Vegas, NV, 2012, pp. 84-92, doi: 10.1109/MASS.2012.6502505.
[3].	P. Huang, Z. Kang, C. Liu and F. Lin, "ACO-based path planning scheme in RWSN," 2016 10th International Conference on Software, Knowledge, Information Management & Applications (SKIMA), Chengdu, 2016, pp. 237-242, doi: 10.1109/SKIMA.2016.7916226.
[4].	A. Tomar, A. Kaswan and P. K. Jana, "On-Demand Energy Provisioning in Wireless Sensor Networks with Capacity-Constrained Mobile Chargers," 2018 Eleventh International Conference on Contemporary Computing (IC3), Noida, 2018, pp. 1-6, doi: 10.1109/IC3.2018.8530654.
[5].	X. Gu, J. Peng, X. Zhang, K. Liu and Y. Cheng, "A Density-Based Clustering Approach for Optimal Energy Replenishment in WRSNs," 2017 IEEE International Symposium on Parallel and Distributed Processing with Applications and 2017 IEEE International Conference on Ubiquitous Computing and Communications (ISPA/IUCC), Guangzhou, 2017, pp. 1018-1023, doi: 10.1109/ISPA/IUCC.2017.00155.
[6].	S. A. Chowdhury and A. Benslimane, "Recharging of Wireless Sensor Network using KMEC with dynamic active zone strategy," 2016 IEEE 12th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob), New York, NY, 2016, pp. 1-6, doi: 10.1109/WiMOB.2016.7763236.
[7].	L. He, Y. Gu, J. Pan and T. Zhu, "On-demand Charging in Wireless Sensor Networks: Theories and Applications," 2013 IEEE 10th International Conference on Mobile Ad-Hoc and Sensor Systems, Hangzhou, 2013, pp. 28-36, doi: 10.1109/MASS.2013.51.
[8].	Y. Feng, L. Guo, X. Fu and N. Liu, "Efficient Mobile Energy Replenishment Scheme Based on Hybrid Mode for Wireless Rechargeable Sensor Networks," in IEEE Sensors Journal, vol. 19, no. 21, pp. 10131-10143, 1 Nov.1, 2019, doi: 10.1109/JSEN.2019.2928169.
[9].	C. Yang, K. Shih and S. Chang, "A Priority-Based Energy Replenishment Scheme for Wireless Rechargeable Sensor Networks," 2017 31st International Conference on Advanced Information Networking and Applications Workshops (WAINA), Taipei, 2017, pp. 547-552, doi: 10.1109/WAINA.2017.112.
[10].	G. Han, Z. Li, J. Jiang, L. Shu and W. Zhang, "MCRA: A Multi-Charger Cooperation Recharging Algorithm Based on Area Division for WSNs," in IEEE Access, vol. 5, pp. 15380-15389, 2017, doi: 10.1109/ACCESS.2017.2727041.
[11].	X. Wang, Q. Zhou, C. Qu, G. Chen and J. Xia, "Location Updating Scheme of Sink Node Based on Topology Balance and Reinforcement Learning in WSN," in IEEE Access, vol. 7, pp. 100066-100080, 2019, doi: 10.1109/ACCESS.2019.2929756.