基于LS-SVM的半圆拱形巷道无线信道建模与预测

发布时间：2018-04-03 12:30

本文选题：矿井巷道　切入点：电磁波传播　出处：《西安科技大学》2017年硕士论文

【摘要】：近年来随着科技发展,矿井安全事故发生率已经相对减少,但是井下的安全生产问题和矿难救援的有效性问题仍然严重,亟需一个可靠、灵敏的井下无线通信系统。但是,由于井下环境复杂,无线电波在其中的传输将受到多种因素的影响,导致井下电波传播衰落复杂,无线信道预测缺乏准确性,因而限制了无线通信系统在井下的传输距离。针对此问题,本文在分析无线电波在井下传播特性理论的基础上,使用最小二乘支持向量机(LS-SVM)建立井下无线信道的预测模型,为研究和设计井下无线通信系统提供指导意义。首先介绍了支持向量机的相关理论基础,分析了电磁波在井下的固有传播模式,重点研究了天线的工作频率、收发天线距离、煤层电参数、巷道壁粗糙和倾斜以及巷道截面对电磁波衰减的影响。以半圆拱形巷道为研究对象,分别对巷道中的路径损耗和多径衰落进行建模和预测。针对路径损耗,将以上影响电磁波传播的因子作为输入变量,以路径损耗值作为预测输出,建立基于LS-SVM的预测模型。在模型训练时,利用遗传算法对LS-SVM中的惩罚因子和核函数参数的选择进行优化。通过和小波神经网络以及基于网格搜索的LS-SVM方法对比,验证算法的有效性。针对多径衰落,以Nakagami衰落信道为研究对象。对Nakagami衰落信道进行仿真得到信道衰落系数变化的时间序列,首先在数据预处理阶段采用主成分分析法处理数据,然后利用相空间重构对衰落系数进行重构,在此基础上建立多径衰落的LS-SVM预测模型。通过仿真实验可以看出,基于LS-SVM的井下巷道预测模型能够较好地预测信道的变化情况。对比传统的预测算法,使用改进的最小二乘支持向量机算法的预测数据和实际数据之间平均相对误差、均方根误差和相关系数三个参数都有所改善,提高了井下巷道无线信号的预测的精度,说明利用支持向量机技术建立井下电波传播模型是一个可行的途径。
[Abstract]:In recent years, with the development of science and technology, the incidence of mine safety accidents has been relatively reduced, but the problems of underground safety production and the effectiveness of mine rescue are still serious, and a reliable and sensitive underground wireless communication system is urgently needed.However, due to the complex underground environment, the transmission of radio waves will be affected by many factors, which leads to the complex downhole radio wave propagation and the lack of accuracy in radio channel prediction.Therefore, the transmission distance of wireless communication system in underground is limited.In order to solve this problem, on the basis of analyzing the theory of radio wave propagation in underground, this paper uses least square support vector machine (LS-SVM) to establish the prediction model of underground radio channel, which provides guidance for the research and design of underground radio communication system.Firstly, the related theory of support vector machine is introduced, and the inherent propagation mode of electromagnetic wave is analyzed. The working frequency of antenna, the distance of receiving and receiving antenna, the electric parameters of coal seam are studied.The influence of roadway wall roughness and inclination and roadway cross section on electromagnetic wave attenuation.Taking the semi-circular arch roadway as the research object, the path loss and multipath fading in the tunnel are modeled and predicted respectively.According to the path loss, the factors that affect the propagation of electromagnetic wave are taken as input variables, and the path loss value is taken as the predictive output. The prediction model based on LS-SVM is established.In model training, genetic algorithm is used to optimize the selection of penalty factors and kernel function parameters in LS-SVM.Compared with wavelet neural network and LS-SVM method based on grid search, the validity of the algorithm is verified.For multipath fading, the Nakagami fading channel is taken as the research object.The time series of channel fading coefficient change are obtained by simulating the Nakagami fading channel. Firstly, the principal component analysis method is used to process the data in the data preprocessing stage, and then the fading coefficient is reconstructed by phase space reconstruction.On this basis, the LS-SVM prediction model of multipath fading is established.The simulation results show that the tunnel prediction model based on LS-SVM can well predict the channel change.Compared with the traditional prediction algorithm, the average relative error, root mean square error and correlation coefficient between the prediction data and the actual data of the improved least squares support vector machine algorithm are improved.The accuracy of radio signal prediction in underground roadway is improved. It is proved that it is a feasible way to establish underground radio wave propagation model by using support vector machine (SVM) technology.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN92;TP18

【参考文献】