基于SVM的网络定位算法的仿真设计

The Positioning Technology of Sensor Network Based on Support Vector Machine

定位问题是无线传感器网络的基本问题,对监控整个平台起到了支撑作用。目前,由于传感器节点受自身计算能力、功耗、通信能力等条件的制约,还没有一个通用的定位方法。因此,新的传感器网络定位方法的研究具有十分重要的意义。重点研究了利用Matlab软件分析支持向量机(SVM)算法中的核函数在传感器定位中的应用。采用内核方法针对传感器网络的粗粒度的定位问题,解决统计学习理论的内核方法模式识别问题,并用算法评估模拟传感器网络。

The positioning problem is the basic problem of wireless sensor networks,and plays a supporting role in monitoring the entire platform.At present,there is no universal positioning method for sensor nodes by the constraints of their own computing power,power consumption,communication capabilities,etc..Therefore,the research of new sensor network localization methods is of great significance.This paper focuses on the application of MATLAB software to analyze the kernel function application in the support vector machine(SVM)algorithm in sensor positioning.The kernel method is applied to solve the coarsegrained positioning problem of the sensor network,and the kernel method pattern recognition problem of statistical learning theory.The algorithm evaluated simulation sensor network is used.