基于支持向量机的无线传感器网络节点定位算法

发布时间：2018-06-13 23:42

本文选题：无线传感器网络 + 节点定位　；参考：《计算机研究与发展》2014年11期

【摘要】：机器学习是利用经验来改善自身性能的一种学习方法,而支持向量机(support vector machine,SVM)作为机器学习中的一种新模式,在解决小样本、非线性及高维模式识别等方面有着其特有的优势.基于支持向量机的节点定位算法利用机器学习算法的特性,实现无线传感网络节点定位.其基本思路是将网络区域划分为若干个等分的小格,每一小格代表机器学习算法中一个确定的类别,机器学习算法在学习了已知的信标节点对应的类别后,对未知节点所处位置进行分类,从而进一步确定未知节点的位置坐标.仿真实验表明,"一对一"节点定位算法有较高的定位精度,对测距误差的容忍性较好,同时对信标节点的比例要求并不高,比较适合用于信标节点稀疏的网络环境中;而"决策树"节点定位算法受覆盖空洞的影响并不大,比较适合应用于节点分布不均匀或者存在覆盖空洞的网络环境中.
[Abstract]:Machine learning is a learning method using experience to improve its performance. As a new model in machine learning, support vector machine (SVM) has its unique advantages in solving small samples, nonlinear and high dimensional pattern recognition. The node localization algorithm based on support vector machine (SVM) realizes node localization in wireless sensor networks by using the characteristics of machine learning algorithm. The basic idea of the algorithm is to divide the network region into a number of equally divided cubes, each of which represents a certain category in the machine learning algorithm, and the machine learning algorithm after learning the categories corresponding to the known beacon nodes. The location of unknown nodes is classified to determine the coordinates of unknown nodes. The simulation results show that the "one-to-one" node localization algorithm has higher localization accuracy, better tolerance to ranging errors, and less requirement for the proportion of beacon nodes, so it is more suitable for the sparse network environment of beacon nodes. The "decision tree" node location algorithm is not greatly affected by the coverage hole, so it is more suitable to be used in the network environment where the node distribution is not uniform or the overlay hole exists.
【作者单位】：浙江工业大学计算机科学与技术学院;
【基金】：国家自然科学基金项目(61379023) 浙江省自然科学基金项目(LY12F02036) 浙江省科技计划基金项目(2011C21014)
【分类号】：TP181;TP212.9;TN929.5

【参考文献】