无线传感器网络中的盲检测算法研究
发布时间：2018-07-20 20:46
【摘要】：无线传感器网络(WSN)是由洒布在目标监测范围内大量微型智能传感器节点组成,具有功耗低、体积小、成本低、分布式以及自组织等特点。由于网络中传感器节点携带能量有限,并且在大多数情况下节点不能充电或更换,因此低能耗成为无线传感器网络设计的主要原则之一,促使不使用训练序列的盲均衡、盲检测技术成为无线传感器网络信号检测技术的研究方向之一。在项目组前期成果基础上,本论文主要创新工作如下:(1)本文在研究经典基于二阶统计量的线性预测算法及其改进算法基础上,结合递归最小二乘算法,提出了一种RLS-MSPA算法,该算法不仅避免了线性预测算法中的输出相关矩阵求逆问题;并且实验表明,相比较于线性预测算法,该算法具有更优的盲检测性能。(2)现有文献给出的无线传感器网络盲检测系统的缺陷之一就是速度过慢,簇内基准传感器节点盲检测速度对整个系统至关重要,为提高系统计算速度,结合提出的RLS-MSPA算法,本文构建了一种新的RLS-MSPA虚拟MIMO无线传感器网络盲检测模型:将RLS-MSPA算法应用于簇内基准传感器节点和簇外sink节点,簇内其它非基准传感器节点则采用互相关算法恢复原信号,从系统模型整体角度出发,通过簇内和簇外两层信号盲检测恢复无线传感器网络中各节点的发送数据。(3)由于RLS-MSPA是基于二阶统计量的盲检测算法,对含公零点的信道不适用,为进一步提高系统对信道适应能力和盲检测性能,本文通过引入改进的蚁群算法SSAV-QACO应用于基准传感器节点信号盲检测,提出了改进的蚁群算法SSAV-QACO虚拟MIMO无线传感器网络盲检测系统。相比较于基准传感器节点采用的RLS-MSPA算法,使用SSAV-QACO算法能够获得更优的盲检测性能,并且能适用于更多传输信道,但时间和空间复杂度都远高于RLS-MSPA算法。(4)为权衡系统的信道适应能力、复杂度和性能,结合超混沌预编码提出了DS-NSCNN超混沌预编码虚拟MIMO无线传感网盲检测系统,相比较于RLS-MSPA虚拟MIMO无线传感网盲检测系统,该系统能适应多数信道;相比较于SSAV-QACO虚拟MIMO无线传感网盲检测系统,该系统具有较快的收敛速度和较低的复杂度;并且性能较二者都有所提升。
[Abstract]:Wireless sensor network (WSN) is composed of a large number of micro-smart sensor nodes which are sprinkled in the monitoring range of target. WSN has the characteristics of low power consumption, small size, low cost, distributed and self-organization. Because of the limited energy carried by sensor nodes and the fact that nodes can not be recharged or replaced in most cases, low energy consumption becomes one of the main principles of wireless sensor network design, and promotes blind equalization without training sequence. Blind detection technology has become one of the research directions of wireless sensor network signal detection technology. On the basis of the previous achievements of the project team, the main innovations of this paper are as follows: (1) based on the study of the classical linear prediction algorithm based on second-order statistics and its improved algorithm, a new RLS-MSPA algorithm is proposed based on the recursive least squares algorithm. The algorithm not only avoids the inverse problem of the output correlation matrix in the linear prediction algorithm, but also shows that compared with the linear prediction algorithm, The algorithm has better blind detection performance. (2) one of the shortcomings of the blind detection system in wireless sensor networks is that the speed is too slow. The blind detection speed of the reference sensor nodes in the cluster is very important to the whole system. In order to improve the computing speed of the system, a new blind detection model for RLS-MSPA virtual wireless sensor networks is proposed. The RLS-MSPA algorithm is applied to the intra-cluster reference sensor nodes and the out-of-cluster sink nodes. Other non-reference sensor nodes in the cluster use cross-correlation algorithm to recover the original signal, and proceed from the system model as a whole. The transmit data of each node in wireless sensor network is recovered by blind detection of two layers of signals inside and outside the cluster. (3) because RLS-MSPA is a blind detection algorithm based on second-order statistics, it is not suitable for channels with common zeros. In order to improve the channel adaptability and blind detection performance of the system, an improved ant colony algorithm (SSAV-QACO) is introduced to blind signal detection of reference sensor nodes. An improved ant colony algorithm (SSAV-QACO) blind detection system for wireless sensor networks is proposed. Compared with the RLS-MSPA algorithm used in the reference sensor node, the SSAV-QACO algorithm can achieve better blind detection performance and can be applied to more transmission channels. But the complexity of time and space is much higher than that of RLS-MSPA algorithm. (4) in order to balance the channel adaptability, complexity and performance of the system, a DS-NSCNN hyperchaotic precoding virtual MIMO wireless sensor network blind detection system is proposed. Compared with RLS-MSPA virtual MIMO wireless sensor network blind detection system, this system can adapt to most channels, compared with SSAV-QACO virtual MIMO wireless sensor network blind detection system, the system has faster convergence speed and lower complexity. And the performance is improved compared with both.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5
本文编号：2134726
[Abstract]:Wireless sensor network (WSN) is composed of a large number of micro-smart sensor nodes which are sprinkled in the monitoring range of target. WSN has the characteristics of low power consumption, small size, low cost, distributed and self-organization. Because of the limited energy carried by sensor nodes and the fact that nodes can not be recharged or replaced in most cases, low energy consumption becomes one of the main principles of wireless sensor network design, and promotes blind equalization without training sequence. Blind detection technology has become one of the research directions of wireless sensor network signal detection technology. On the basis of the previous achievements of the project team, the main innovations of this paper are as follows: (1) based on the study of the classical linear prediction algorithm based on second-order statistics and its improved algorithm, a new RLS-MSPA algorithm is proposed based on the recursive least squares algorithm. The algorithm not only avoids the inverse problem of the output correlation matrix in the linear prediction algorithm, but also shows that compared with the linear prediction algorithm, The algorithm has better blind detection performance. (2) one of the shortcomings of the blind detection system in wireless sensor networks is that the speed is too slow. The blind detection speed of the reference sensor nodes in the cluster is very important to the whole system. In order to improve the computing speed of the system, a new blind detection model for RLS-MSPA virtual wireless sensor networks is proposed. The RLS-MSPA algorithm is applied to the intra-cluster reference sensor nodes and the out-of-cluster sink nodes. Other non-reference sensor nodes in the cluster use cross-correlation algorithm to recover the original signal, and proceed from the system model as a whole. The transmit data of each node in wireless sensor network is recovered by blind detection of two layers of signals inside and outside the cluster. (3) because RLS-MSPA is a blind detection algorithm based on second-order statistics, it is not suitable for channels with common zeros. In order to improve the channel adaptability and blind detection performance of the system, an improved ant colony algorithm (SSAV-QACO) is introduced to blind signal detection of reference sensor nodes. An improved ant colony algorithm (SSAV-QACO) blind detection system for wireless sensor networks is proposed. Compared with the RLS-MSPA algorithm used in the reference sensor node, the SSAV-QACO algorithm can achieve better blind detection performance and can be applied to more transmission channels. But the complexity of time and space is much higher than that of RLS-MSPA algorithm. (4) in order to balance the channel adaptability, complexity and performance of the system, a DS-NSCNN hyperchaotic precoding virtual MIMO wireless sensor network blind detection system is proposed. Compared with RLS-MSPA virtual MIMO wireless sensor network blind detection system, this system can adapt to most channels, compared with SSAV-QACO virtual MIMO wireless sensor network blind detection system, the system has faster convergence speed and lower complexity. And the performance is improved compared with both.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5
【参考文献】
相关期刊论文 前10条
1 陈咨皓;;WSN在城市智能交通中的应用[J];信息通信;2015年03期
2 张振洲;于舒娟;张昀;;基于无线传感器网络虚拟MIMO系统盲均衡[J];计算机技术与发展;2014年03期
3 冯迪;于舒娟;张昀;;一种改进激活函数的Hopfield盲检测算法[J];计算机技术与发展;2012年12期
4 常超;鲜晓东;胡颖;;基于WSN的精准农业远程环境监测系统设计[J];传感技术学报;2011年06期
5 党月芳;;无线传感器网络在军事领域中的应用研究[J];信息通信;2011年03期
6 王兴林;李茂军;;基于改进量子遗传算法的Flow-Shop调度求解[J];计算技术与自动化;2010年03期
7 蔡蓓蓓;张兴华;;混合量子遗传算法及其在VRP中的应用[J];计算机仿真;2010年07期
8 彭景斌;叶进宝;王雪娇;;暂态混沌神经网络及其在优化问题中的应用研究[J];现代电子技术;2009年04期
9 李长庚;周家令;孙克辉;盛利元;;四种数字混沌扩频序列的平衡性分析[J];计算机应用;2008年01期
10 张志涌,王俊;信道盲辨识的新方法[J];电子学报;2003年S1期
相关博士学位论文 前1条
1 胡小兵;蚁群优化原理、理论及其应用研究[D];重庆大学;2004年
相关硕士学位论文 前7条
1 季奎明;改进的Hopfield型神经网络盲检测算法研究[D];南京邮电大学;2016年
2 胡蓉;基于分簇虚拟MIMO无线传感网盲处理系统研究[D];南京邮电大学;2015年
3 李瑞翔;面向精细农业WSN中的信号盲处理问题及研究[D];南京邮电大学;2014年
4 张振洲;基于分簇虚拟MIMO无线传感器网络盲检测系统[D];南京邮电大学;2014年
5 边玮;一种基于动态分簇的无线传感器网络跟踪算法研究[D];吉林大学;2013年
6 蔡晴红;量子遗传算法在盲检测中的研究与应用[D];南京邮电大学;2013年
7 冯迪;基于改进型Hopfield神经网络的盲检测新算法研究[D];南京邮电大学;2013年
，本文编号：2134726
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2134726.html
