基于Q-Learning反馈机制的无线传感网络通信节点自愈算法  

Self-Healing Algorithm of Wireless Sensor Network Communication Node Based on Q-Learning Feedback Mechanism

作　　者：杨惠 YANG Hui(School of Media Engineering,Lanzhou University of Arts and Science,Lanzhou Gansu 730000,China)

机构地区：[1]兰州文理学院传媒工程学院,甘肃兰州730000

出　　处：《传感技术学报》2022年第7期974-979,共6页Chinese Journal of Sensors and Actuators

基　　金：甘肃省教育厅创新基金项目(2022A-171);甘肃省自然科学基金项目(1606RJZA181)。

摘　　要：针对目前无线网络通信节点自愈能力差,以及自愈后网络流量出口带宽低的问题,提出基于Q-learning反馈机制的无线传感网络通信节点自愈算法。通过计算网路节点的RSSI值建立节点衰减模型,通过质心算法完成节点定位;应用Q-learning学习算法获取链路选取策略,完成节点传输过程路径时延、吞吐量以及丢包率的计算,建立网络节点模型提取链路反馈机制,利用Q-learning学习算法进行迭代计算,实现无线传感网络的通信节点自愈。仿真分析表明,运用该算法自愈网络通信节点时,当检测次数为100时,检测出的节点自愈数量为280个,节点拓扑移动距离平均值为175 m,网络流量出口带宽平均值为550 Mbyte/s,证明该算法的节点自愈能力高。Aiming at the problem of poor self-healing ability of wireless network communication nodes and low bandwidth of network traffic after self-healing,a self-healing algorithm for wireless sensor network communication nodes based on Q-learning feedback mechanism is proposed.The node attenuation model is established by calculating the RSSI value of the network node,and the node positioning is completed by using the centroid algorithm;the Q-learning algorithm is used to obtain the link selection strategy,and the calculation of path delay,throughput and packet loss rate of the node transmission process is completed.The network node model is established to extract the link feedback mechanism,and the Q-learning algorithm is used to perform iterative calculations to realize the self-healing of the communication nodes of the wireless sensor network.Simulation analysis shows that through applying this algorithm to self-heal network communication nodes,when the number of detections is 100,the number of self-healing nodes detected is 280,the average moving distance of the node topology is 175 m,and the average network traffic outlet bandwidth is 550 Mbyte/s.This proves that the node self-healing ability of this algorithm is high.

关 键 词：无线传感网络 通信节点自愈 Q-learning学习算法 节点定位 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]