改進型模式切換網格網絡架構整合粒子群演算法和模糊邏輯控制與田口方法應用於無線感測網路

Improved Mode-Switched Grid-based Architecture Networks Combined PSO and FLC with Taguchi Methods in Wireless Sensor Networks

蕭冠文(Gun-Wen, Xiao)

指導教授 ： 陳永隆

國立臺中科技大學/資訊與流通學院/資訊工程系碩士班/碩士(2019年)

摘要

在無線感測網路中，無線感測節點能量消耗是很重要的議題之一。無線感測網路主要是由大量的無線感測節點所構成，其中感測節點非均勻分布。無線感測節點大部份都是使用電池來提供能量，而通常無線感測節點會佈署在人們不易到達的地方，使得人們無法輕易更換電池，造成感測節點死亡無法繼續蒐集和傳遞資料，導致整個無線感測網路的生命週期下降。因此有效地減少節點能量消耗並提高傳輸效率以及延長網路生命週期是非常重要的。本論文中，我們參考MSGR方法的網格架構重新設計提出三個方法，分別為: IMSGOP-FLCT、IMSHGOR-FLCPSOT和IMSPOR-FLCPSOT。在本論文所提出之IMSGOP-FLCT方法的網格形成階段，無線感測場域將切成數個均勻大小的網格，並在每個網格利用模糊邏輯控制結合田口方法選出具有最佳剩餘能量及位置的節點作為網格頭，接著進入網格睡眠排程階段，計算網格的睡眠排程，使網格以固定時間輪流休眠以此節省能量消耗。在資料傳輸階段，來源感測節點將資訊傳遞給鄰近網格頭，若來源感測節點所屬的網格處於睡眠狀態則使用模糊邏輯控制結合田口方法找尋最佳傳輸路徑。進一步，為了改善IMSGOP-FLCT方法的資料傳輸效率，本論文提出的IMSHGOR-FLCPSOT方法在網格頭形成階段仍然使用模糊邏輯控制結合田口方法選擇最佳的網格頭，在傳輸階段，來源端將計算與接收端的距離是否達到門檻，若達則使用粒子群演算法與田口方法尋找最佳轉傳節點，以此提高資料傳輸速率並降低傳輸延遲。最後，為了避免低節點密度的網格因傳遞資料的能量耗損，本論文提出的IMSPOR-FLCPSOT方法在網格形成階段使用模糊C聚類方法選出具有最佳能量及位置的節點作為網格頭，並在網格形成階段後計算每個網格的節點密度、鄰居網格與平均網格密度，當網格密度達到分割的門檻值則將該網格使用模糊C聚類方法分為兩群並找出網格頭，以此可使粒子群演算法在搜尋轉傳節點時能找到具有更佳位置的節點做為轉傳節點來轉送資料。從模擬結果證明，我們提出的IMSGOP-FLCT方法及IMSHGOR-FLCPSOT方法與IMSPOR-FLCPSOT方法比起EAGER方法與MSGR方法擁有更佳地網格頭尋找效率，減少了資料跳傳次數並節省了探測路由所花費的能量，在傳輸資料方面有更低的資料傳遞延遲，拓展了整體無線場測網路資料傳輸生命週期。

關鍵字

無線感測網路 ； 模糊C聚類 ； 模糊邏輯控制 ； 粒子群演算法 ； Grid

並列摘要

Energy consumption of wireless sensor node is one of the most important issues in wireless sensor networks (WSNs). The wireless sensor networks is mainly composed of a large number of wireless sensor nodes, wherein the sensor nodes are non-uniformly distributed. Most of the wireless sensor nodes use batteries to provide energy. Usually, the wireless sensor nodes are deployed in places that are difficult for people to reach, so that people cannot easily replace the batteries, causing the death of the sensor nodes to continue collecting and transmitting data, and it leads to a decline in the life cycle of the entire wireless sensor networks. Therefore, to effectively reduce node energy consumption, improve transmission efficiency, and extend the network life cycle is very important. In this thesis, we propose three methods for redesigning the grid architecture: An improved mode-switched grid-based architecture with optimal path combined fuzzy logic control and Taguchi methods for wireless sensor Networks (IMSGOP-FLCT), An improved mode-switched hierarchy grid-based architecture with optimal relay combined fuzzy logic control and particle swarm optimization with Taguchi methods for wireless sensor networks (IMSHGOR-FLCPSOT) and An improved mode-switched partition-based architecture With optimal relay combined fuzzy logic control and particle swarm optimization with Taguchi methods for wireless sensor networks (IMSPOR-FLCPSOT). In the grid formation phase of the IMSGOP-FLCT method proposed in this thesis, the wireless sensor field will be cut into several uniform-sized grids, and the best energy can be selected by using the FLC method with Taguchi method to allocate the GH (grid head, GH) in each grid. The node of the position is the GH, and then enters the grid sleep scheduling phase, and the sleep schedule of each grid is calculated, so that the grid runs or sleeps in a fixed time to save energy consumption. In the data transmission phase, the source sensor node transmits the information data to the neighboring GH. If the grid to which the source sensor node belongs is in a sleep state, the FLC method is used in conjunction with the Taguchi method to find the optimal transmission path. In order to improve the data transmission efficiency of the IMSGOP-FLCT method, the IMSHGOR-FLCPSOT method proposed in this thesis will calculate whether the distance from the receiving end reaches the threshold at the transmission phase. If the threshold is reached, the PSOT (Particle Swarm Optimization algorithm with Taguchi method, PSOT) are used. The POST finds the best transfer node to increase the data transmission rate and reduce the transmission delay. In order to avoid the energy loss grid of the data transmitted by the low- sensor node density, the IMSPOR-FLCPSOT method proposed in this thesis uses the FCM method to select the node with the best energy and position as the mesh header in the mesh formation phase. After the grid formation phase, calculate the shortest distance between each grid, the grid density, the neighbor grid and the average grid density. When the grid density reaches the threshold of the segmentation, the grid is divided into two groups using the FCM method. The grid header is used, so that the PSOT algorithm can find a node with a better position as a transit node to transfer data when searching for the forwarding node. The proposed IMSGOP-FLCT method and the IMSHGOR-FLCPSOT method and the IMSPOR-FLCPSOT method have better grid head finding efficiency than the EAGER method and the MSGR method, reducing the number of data hopping and saving the energy spent on detecting routes. There is a lower data transfer delay in the transmission of data, expanding the overall infinite field measurement network life.

並列關鍵字

WSNs ； particle swarm optimization ； fuzzy C-means ； fuzzy logic control ； grid

參考文獻

[1] M. Tolani, Sunny, R. K. Singh, K. Shubham, and R. Kumar, “Two-layer optimized railway monitoring system using Wi-Fi and ZigBee interfaced wireless sensor network,” IEEE Sensors J., vol. 17, no. 7, pp. 2241-2248, Apr. 2017.

[2] F. Tian, X. Chen, S. Liu, X. Yuan, D. Li, X. Zhang, and Z. Yang, “Secrecy rate optimization in wireless multi-hop full duplex networks,” IEEE Access, vol. 6, pp. 5695-5704, Jan. 2018.

[3] F. Tian, X. Chen, S. Liu, K. Wang, X. Yuan, and Z. Yang, “On full duplex scheduling for energy efficiency maximization in multi-hop wireless networks,” IEEE Access, vol. 6, pp. 2604-2614, Dec. 2017.

[4] N. Kumar and D. P. Vidyarthi, “A green routing algorithm for IoT-enabled software defined wireless sensor network,” IEEE Sensors J., vol. 18, no. 22, pp. 9449-9460, Sep. 2018.

[5] H. Huang, H. Yin, G. Min, J. Zhang, Y. Wu, and X. Zhang, “Energy-aware dual-path geographic routing to bypass routing holes in wireless sensor networks,” IEEE Trans. Mobile Comput., vol. 17, no. 6, pp. 1339-1352, June 2018.