ISSN 1000-1239 CN 11-1777/TP

计算机研究与发展 ›› 2014, Vol. 51 ›› Issue (11): 2382-2392.doi: 10.7544/issn1000-1239.2014.20131079

• 网络技术 • 上一篇    下一篇

无线传感器网络通信负载状态识别方法的研究

赵泽1,尚鹏飞1,2,刘强1,崔莉1   

  1. 1(中国科学院计算技术研究所 北京 100190);2(中国科学院大学计算机与控制工程学院 北京 100190) (zhaoze@ict.ac.cn)
  • 出版日期: 2014-11-01
  • 基金资助: 
    基金项目:国家“九七三”重点基础研究发展计划基金项目(2011CB302803);国家自然科学基金项目(61202412)

Identification of Communication State for Wireless Sensor Networks

Zhao Ze1, Shang Pengfei1,2, Liu Qiang1, Cui Li1   

  1. 1(Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190); 2(School of Computer and Control Engineering, University of Chinese Academy of Sciences, Beijing 100190)
  • Online: 2014-11-01

摘要: 在无线传感器网络数据传输过程中,网络的通信状态是一个重要的研究点.当网络吞吐率过载时,需要调整网络通信策略以保证传感网的整体性能,因此首先需要对网络的传输状态进行识别.设计并实现了一种能够实时在线识别传感器网络数据通信状态的机制,能够对网络正常、过载的传输状态进行准确判断,以支持网络通信策略的调整.首先通过实验获得网络传输过程中的基础性能参数,并对这些参数与网络通信负载状态的相关性进行分析,选择吞吐率作为网络负载状态的分类标准,建立基于支持向量机模型(support vector machines, SVM)的网络通信负载状态识别模型,并根据实测数据确定该模型在模拟环境中的参数.实际测试表明,该判别模型的准确率可达到91.28%,能够对网络通信负载状态进行有效识别.

关键词: 无线传感器网络, 负载状态, 分类识别, NS2, 支持向量机

Abstract: In the process of wireless sensor network communication, the communication state of network is a quite important research point. Network communication strategy needs to be adjusted to ensure the overall performance when the network throughput overloads. Therefore, we need to identify the state of network transmission as the guidance of network communication strategy adjustment firstly. In this paper, we design and implement a real-time online mechanism to identify the communication state of the sensor networks, which can accurately determine whether the network transmission state is normal or overload, to support the adjustment of communication strategy. We obtain the basic performance parameters of the process of transmission by experiments, and analyze the correlation between these parameters and the communication state. Then, we select throughput as the classification criteria of the network communication state, and establish the identification model of communication state based on the support vector machines (SVM), and obtain the parameters of the model in the simulated environment according to the experimental data. Evaluation result shows that the accuracy of the identification model can be achieved up to 91.28%, which can effectively identify the communication state for wireless sensor networks.

Key words: wireless sensor network (WSN), communication state, classification and identification, NS2, support vector machines

中图分类号: