基于BP神经网络的LTE上行干扰定位算法
详细信息 查看官网全文
- 英文论文题名：LTE uplink interference identification algorithm based on BP neural network
- 论文作者：李翔明 ; 李冰 ; 高超云 ; 孙磊
- 英文论文作者：LI Xiang-ming ; LI Bing ; GAO Chao-yun ; SUN Lei ; China mobile group yunnan co. ; LTD
- 年：2016
- 作者机构：中国移动通信集团云南有限公司;
- 论文关键词：LTE ; 上行干扰 ; BP神经网络 ; 干扰定位
- 英文论文关键词：LTE ; uplink interference ; BP artificial neural network ; interference identification
- 会议召开时间：2016-09-06
- 会议录名称：第六届云南省科协学术年会暨红河流域发展论坛论文集——专题二：滇南中心智慧城市建设
- 语种：中文
- 分类号：TN929.5
- 学会代码：SSAD
- 会议名称：第六届云南省科协学术年会暨红河流域发展论坛
- 会议地点：中国云南红河
- 主办单位：云南省科学技术协会、红河州人民政府、中国通信学会
- 学会名称：云南省机械工程学会
- 页数：7
- 文件大小：756k
- 原文格式：D
- 会议级别：地方
摘要
TD-LTE系统同频组网,现网中上行干扰是系统干扰的主要来源。为了实现对TD-LTE无线小区受到的上行干扰类型进行自动定位。以BP人工神经网络算法为基础,选取900个现网干扰小区PRB干扰电平数据作为建模样本,采样200个现存干扰小区作为检验样本,建立了LTE上行干扰定位模型,其结果是干扰的定位准确率达到90%。结果表明,BP神经网络本构关系模型具有很高的预测精度。经过对节点数和训练次数的优化,可以很好地对现网存在的系统内干扰、阻塞干扰、谐波干扰、杂散干扰进行分类。本算法有助于提升干扰定位工作效率,具有较高的应用价值。
TD-LTE system is intra-frequency network. The uplink interference is the major source of system interference. In order to automatically identify the interference type of the E-UTRAN cells in LTE system. A method based on BP artificial neural network algorithm is applied. By means of selecting the PRB interference level data of 900 cells as the training sample and sampling the level data of another 200 cells as the testing sample, a LTE uplink interference identification model is designed. Experimental results show that, the constitutive relationship model of the BP neural network makes prediction accurately. By optimizing the number of the hidden layers and the training times, the identification of the intra-system interference, blocking interference, harmonic interference and spurious interference is achieved,with more than 90% accuracy. This method helps to improve the efficiency of interference identification and demonstrates the high application value.
TD-LTE system is intra-frequency network. The uplink interference is the major source of system interference. In order to automatically identify the interference type of the E-UTRAN cells in LTE system. A method based on BP artificial neural network algorithm is applied. By means of selecting the PRB interference level data of 900 cells as the training sample and sampling the level data of another 200 cells as the testing sample, a LTE uplink interference identification model is designed. Experimental results show that, the constitutive relationship model of the BP neural network makes prediction accurately. By optimizing the number of the hidden layers and the training times, the identification of the intra-system interference, blocking interference, harmonic interference and spurious interference is achieved,with more than 90% accuracy. This method helps to improve the efficiency of interference identification and demonstrates the high application value.
引文
[1]李行政,张冬晨,姚文闻,何继伟.一种TD-LTE系统上行干扰三维分析方法[J].电信工程技术与标准化,2016,29(6)
[2]王浩全.基于BP神经网络提高伪装目标识别概率的研究[J].光谱学与光谱分析,2010,30(12)
[3]张军.TD-LTE宏蜂窝上行干扰优化思路[J].通讯世界,2015,(19)
[4]孙宇,曾卫东,等.基于BP神经网络Ti600合金本构关系模型的建立[J].稀有金属材料与工程,2011,40(2)
[5]余凡,赵英时,李海涛.基于遗传BP神经网络的主被动遥感协同反演土壤水分[J].红外与毫米波学报,2012,31(3)
[6]张宝菊,雷晴.基于BP神经网络的人体血液中红细胞浓度无创检测[J].光谱学与光谱分析,2012,32(9)
[2]王浩全.基于BP神经网络提高伪装目标识别概率的研究[J].光谱学与光谱分析,2010,30(12)
[3]张军.TD-LTE宏蜂窝上行干扰优化思路[J].通讯世界,2015,(19)
[4]孙宇,曾卫东,等.基于BP神经网络Ti600合金本构关系模型的建立[J].稀有金属材料与工程,2011,40(2)
[5]余凡,赵英时,李海涛.基于遗传BP神经网络的主被动遥感协同反演土壤水分[J].红外与毫米波学报,2012,31(3)
[6]张宝菊,雷晴.基于BP神经网络的人体血液中红细胞浓度无创检测[J].光谱学与光谱分析,2012,32(9)