Two-Tier Data Dissemination Protocols with Q-Learning for Wireless Sensor Networks

• Main Authors: HSU, WEI-JUNG, 徐巍榮
• Other Authors: WANG, NENG-CHUNG
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/63b3ry

碩士 === 國立聯合大學 === 資訊工程學系碩士班 === 106 === 　　A wireless sensor network (WSN) consists of a great number of sensor nodes. Sensor nodes itself has the ability to wireless communicate, detect and process data. Through the sensor nodes, we can detect the surrounding environment of sensor nodes and transmit it to the sink with wireless communication. In a wireless sensor network, the energy of sensor node is limited. Therefore, designing an energy efficient data transmission scheme of WSNs to reduce energy consumption is a very important issue. In this thesis, we first propose a two-tier data dissemination protocol with Q-learning for wireless sensor networks, called TTDD-QL. In the proposed scheme, the source uses Q-learning to find an energy efficient data dissemination path from the source to the sink. In TTDD-QL, the reward is given by the distance between the dissemination node and the sink. In each iteration, the proposed scheme will update the values of Q table (Q values). After multiple iterations for learning, Q values are converged, then a data dissemination path can be found according to the Q values. We also propose an advance TTDD-QL data dissemination method, called TTDD-QL-A. The proposed TTDD-QL-A is similar to TTTD-QL, the reward is determined not only by the distance between the dissemination node and the sink but also by the current remaining energy of the dissemination node. Simulation result shows that TTDD-QL and TTDD-QL-A can reduce the energy consumption of sensor nodes and extend the lifetime of the WSN.