摘要
针对无线网络室内位置指纹定位中存在定位精度低,跳跃性较大的问题,提出一种烟花算法优化支持向量机的室内定位模型,对多次采集到的接收信号强度进行高斯滤波去除奇异值,通过烟花算法优化SVM的参数,建立室内无线定位优化模型.实验对比证明,烟花算法比粒子群算法更能提高SVM的优化速率及室内无线网络定位精度.
Aiming at the problem of low positioning accuracy and large span of wireless network on the indoor fingerprint location,we presented a new indoor positioning model based on modified Support Vector Machine（SVM）.Firstly,in order to remove the singular value,we use Gauss Filter to process the received the signal strength values that are collected several times.Secondly,the parameters of support vector machine are optimized by Fireworks Algorithm.Finally,we establish the indoor wireless location model with FWA-SVM.Through experimental comparison,we prove that the fireworks algorithm is more accurate than the Particle Swarm Optimization（PSO） algorithm in the SVM optimization rate and positioning accuracy.
出处
《河北工业大学学报》
CAS
2016年第6期35-40,共6页
Journal of Hebei University of Technology
基金
天津市应用基础与前沿技术研究计划(13JCQNJC00200
14JCYBJC18500)
河北省高等学校科学技术研究项目(ZD20131097)
河北省自然科学基金(F2015202311)
关键词
支持向量机
烟花算法
室内定位
无线网络
高斯滤波
Support vector machine（SVM）
fireworks algorithm（FWA）
indoor positioning
wireless network
Gauss Filter
