基于置信传播的WSN节点定位方法研究
发布时间：2018-08-31 08:10
【摘要】：在商业、公共服务和军事领域,定位技术是无线网络应用的重要技术之一,不含位置信息的网络数据往往是没有意义的。无线传感器网络技术(Wireless Sensor Network, WSN)作为一种具有功耗低、成本低、时延短、可靠、安全等特点的组网技术,近年来得到迅猛的发展。与此同时,无线传感器网络对自身节点的定位有着广泛的需求,但是目前应用最广的全球定位系统因为室内信号弱、定位精度低、功耗高等原因而无法完全满足无线传感器网络技术节点定位的要求,因此无线传感器网络节点定位技术应运而生。为了实现无线传感器网络节点的高精度定位,本文对无线传感器网络节点定位方法进行了深入的研究。目前,根据无线传感器网络节点定位时是否基于测量把节点定位方法分为基于测量的定位方法和测量无关的定位方法,因为基于测量的定位方法的节点定位精度一般高于测量无关的定位方法,所以本文主要研究基于测量的定位方法。到达时间差(Time Difference of Arrival, TDOA)测距技术无须时间同步、测距精度高,是一种无线传感器网络节点定位领域广泛应用的测距技术。本文提出了一种EC-TDOA (Error-Checking TDOA)测距方法,应用实验测距数据和最小二乘回归方法对TDOA测距进行误差校正,获取更为精确的测距结果。采用概率图模型对基于EC-TDOA的无线传感器网络节点定位进行数学建模,并应用置信传播算法和非参数置信传播算法求解概率图模型,得到最大后验概率和节点位置信息。为了验证定位方法的实际效果,设计并实现了一个无线传感器网络节点定位系统。定位系统采用符合ZigBee协议标准的芯片设计节点,应用.NET平台实现上位机软件,可以对无线传感器节点进行实时定位。根据定位系统采集的节点定位数据,对基于TDOA测距的极大似然估计定位方法、基于TDOA测距的非参数置信传播定位方法、基于EC-TDOA测距的极大似然估计定位方法和基于EC-TDOA测距的非参数置信传播定位方法进行对比实验,从节点定位结果的均方根误差、累积误差等多个指标进行讨论与分析。结果表明,基于EC-TDOA测距的非参数置信传播定位方法可以实现,其最大定位误差为4.3cm,平均定位误差为1.64cm,定位频率为17次/秒,其定位性能指标优于其他三种定位方法。与极大似然估计定位方法相比较,基于EC-TDOA测距的非参数置信传播定位方法从测距环节开始降低测量误差,并在基于测距的定位环节实现了对定位精度的二次优化,最后达到提高定位精度的效果高。基于EC-TDOA测距的非参数置信传播定位方法在定位系统上的应用,表现出了良好实时性、鲁棒性,明显改善了节点测量位置的波动现象,起到了滤波效果。
[Abstract]:In the commercial, public service and military fields, location technology is one of the most important technologies in wireless network applications, and network data without location information is often meaningless. Wireless sensor network (Wireless Sensor Network, WSN) technology, which has the characteristics of low power consumption, low cost, short delay, reliability and security, has been developed rapidly in recent years. At the same time, wireless sensor networks have a wide range of requirements for the location of their own nodes, but the most widely used global positioning system because of the weak indoor signal, low positioning accuracy, Because of the high power consumption, it can not completely meet the requirements of node localization in wireless sensor networks, so the node location technology of wireless sensor networks emerges as the times require. In order to achieve high precision location of nodes in wireless sensor networks, this paper makes a deep research on the localization methods of nodes in wireless sensor networks. At present, according to whether the node location of wireless sensor network is based on measurement or not, the node location method is divided into MEASURMENT based localization method and measurement independent localization method. Because the accuracy of node localization based on measurement is generally higher than that of measurement independent, this paper mainly studies the localization method based on measurement. Time difference of arrival (Time Difference of Arrival, TDOA) ranging technology is widely used in the field of node location in wireless sensor networks because it does not need time synchronization and has high ranging accuracy. In this paper, a EC-TDOA (Error-Checking TDOA) ranging method is proposed, which uses the experimental ranging data and the least square regression method to correct the error of TDOA ranging, and obtains more accurate ranging results. The probabilistic graph model is used to model the node location of wireless sensor networks based on EC-TDOA, and the confidence propagation algorithm and nonparametric confidence propagation algorithm are used to solve the probabilistic graph model. The maximum posterior probability and node location information are obtained. In order to verify the effectiveness of the localization method, a node location system for wireless sensor networks is designed and implemented. The positioning system adopts chip design nodes that conform to ZigBee protocol standard, and uses .NET platform to realize upper computer software, which can locate wireless sensor nodes in real time. According to the node location data collected by the localization system, the maximum likelihood estimation localization method based on TDOA ranging and the nonparametric confidence propagation localization method based on TDOA ranging are studied. The maximum likelihood estimation localization method based on EC-TDOA ranging and the nonparametric confidence propagation localization method based on EC-TDOA ranging are compared. The RMS root error and cumulative error of node localization results are discussed and analyzed. The results show that the nonparametric confidence propagation localization method based on EC-TDOA ranging can be realized. The maximum localization error is 4.3 cm, the average positioning error is 1.64 cm, and the localization frequency is 17 times per second. The localization performance index is superior to the other three localization methods. Compared with the maximum likelihood estimation (MLE) localization method, the nonparametric confidence propagation localization method based on EC-TDOA ranging reduces the measurement error from the ranging link, and realizes the quadratic optimization of the location accuracy in the location link based on the ranging. Finally, the effect of improving positioning accuracy is high. The application of the nonparametric confidence propagation localization method based on EC-TDOA ranging in the localization system shows good real-time robustness and obviously improves the fluctuation phenomenon of node measurement position and plays a filtering effect.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9
本文编号：2214408
[Abstract]:In the commercial, public service and military fields, location technology is one of the most important technologies in wireless network applications, and network data without location information is often meaningless. Wireless sensor network (Wireless Sensor Network, WSN) technology, which has the characteristics of low power consumption, low cost, short delay, reliability and security, has been developed rapidly in recent years. At the same time, wireless sensor networks have a wide range of requirements for the location of their own nodes, but the most widely used global positioning system because of the weak indoor signal, low positioning accuracy, Because of the high power consumption, it can not completely meet the requirements of node localization in wireless sensor networks, so the node location technology of wireless sensor networks emerges as the times require. In order to achieve high precision location of nodes in wireless sensor networks, this paper makes a deep research on the localization methods of nodes in wireless sensor networks. At present, according to whether the node location of wireless sensor network is based on measurement or not, the node location method is divided into MEASURMENT based localization method and measurement independent localization method. Because the accuracy of node localization based on measurement is generally higher than that of measurement independent, this paper mainly studies the localization method based on measurement. Time difference of arrival (Time Difference of Arrival, TDOA) ranging technology is widely used in the field of node location in wireless sensor networks because it does not need time synchronization and has high ranging accuracy. In this paper, a EC-TDOA (Error-Checking TDOA) ranging method is proposed, which uses the experimental ranging data and the least square regression method to correct the error of TDOA ranging, and obtains more accurate ranging results. The probabilistic graph model is used to model the node location of wireless sensor networks based on EC-TDOA, and the confidence propagation algorithm and nonparametric confidence propagation algorithm are used to solve the probabilistic graph model. The maximum posterior probability and node location information are obtained. In order to verify the effectiveness of the localization method, a node location system for wireless sensor networks is designed and implemented. The positioning system adopts chip design nodes that conform to ZigBee protocol standard, and uses .NET platform to realize upper computer software, which can locate wireless sensor nodes in real time. According to the node location data collected by the localization system, the maximum likelihood estimation localization method based on TDOA ranging and the nonparametric confidence propagation localization method based on TDOA ranging are studied. The maximum likelihood estimation localization method based on EC-TDOA ranging and the nonparametric confidence propagation localization method based on EC-TDOA ranging are compared. The RMS root error and cumulative error of node localization results are discussed and analyzed. The results show that the nonparametric confidence propagation localization method based on EC-TDOA ranging can be realized. The maximum localization error is 4.3 cm, the average positioning error is 1.64 cm, and the localization frequency is 17 times per second. The localization performance index is superior to the other three localization methods. Compared with the maximum likelihood estimation (MLE) localization method, the nonparametric confidence propagation localization method based on EC-TDOA ranging reduces the measurement error from the ranging link, and realizes the quadratic optimization of the location accuracy in the location link based on the ranging. Finally, the effect of improving positioning accuracy is high. The application of the nonparametric confidence propagation localization method based on EC-TDOA ranging in the localization system shows good real-time robustness and obviously improves the fluctuation phenomenon of node measurement position and plays a filtering effect.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9
【相似文献】
相关期刊论文 前10条
1 翁志诚;甘小莺;徐友云;;一种结合比特翻转的自适应置信传播算法[J];信息技术;2008年04期
2 贺玉成,杨莉,王新梅,福田明;置信传播译码算法的性能测度[J];电子学报;2002年04期
3 贺玉成,慕建君,王新梅;基于置信传播算法的低密度校验码量化译码设计[J];计算机学报;2003年08期
4 史治平;张忠培;朱南;;基于置信传播译码的DRA码设计[J];电子与信息学报;2008年07期
5 郭春生;张大状;;基于置信传播的视频运动目标检测[J];电路与系统学报;2013年01期
6 高恩婷;顾一清;严建峰;;基于快速置信传播算法的并行主题建模方法研究[J];南通大学学报(自然科学版);2013年01期
7 何秀慧;蒋敏兰;;一种改进的LT码置信传播译码[J];计算机工程与应用;2012年14期
8 李昂,罗汉文,陈强;基于置信传播的LDPC码译码算法[J];计算机工程;2005年20期
9 李广文;酆广增;;基于置信传播和波束搜索的LDPC联合译码算法[J];南京邮电大学学报(自然科学版);2008年05期
10 姜小波;李芳苑;;LDPC码的交替迭代分层置信传播译码[J];电路与系统学报;2013年01期
相关会议论文 前1条
1 李广文;酆广增;;基于置信传播和波束搜索的LDPC联合译码算法[A];2008年中国通信学会无线及移动通信委员会学术年会论文集[C];2008年
相关硕士学位论文 前3条
1 张大状;视频运动目标检测的高效置信传播算法研究[D];杭州电子科技大学;2013年
2 方泽军;基于置信传播算法的视频背景估计研究[D];杭州电子科技大学;2011年
3 房卓群;基于置信传播的WSN节点定位方法研究[D];沈阳建筑大学;2014年
，本文编号：2214408
本文链接：https://www.wllwen.com/kejilunwen/wltx/2214408.html
