基于最优功耗分配的分布式估计策略研究

发布时间：2018-02-23 01:31

本文关键词： 分布式融合 Kalman滤波功耗分配信道估计误差估计融合　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：无线传感器网络(Wireless Sensor Networks, WSNs)作为一门交叉学科,结合了传感器技术、无线通信以及计算机网络等技术,近年来引起了广泛关注,成为了国内外的研究热点,也越来越广泛地应用到众多军事和非军事领域,如定位跟踪、远程遥控、军事侦察等。但是由于传感器自身的一些缺陷,比如能量、带宽以及数据处理能力等,使得数据传输效率以及网络生命周期等受到了限制。为了解决通信带宽以及能量受限,提高信息收集的效率,数据融合(Data Fusion)技术起到了非常大的作用。但是,把数据融合应用到实际中,仍然存在很多挑战：传感器系统的功耗分配问题,信道信息的不确定性等,本文针对目前存在的这些问题展开研究,利用压缩策略,使得系统能在降低传输量的同时达到集中式融合的估计性能；研究了在存在同步误差情况下,系统的最优功耗分配,以及信道估计误差对系统性能的影响。具体的工作包括以下内容：1.针对传感器之间量测噪声相关的情况,我们提出了一种基于线性变换的分布式融合算法。与传统的噪声之间互不相关的假设相比,相关噪声的假设更符合实际情况。我们针对线性融合系统提出了一种基于传感器节点量测值的线性变换的无损准则,并给出了能够达到集中式融合性能的必要条件。另外,我们也给出了当量测矩阵为行满秩和列满秩时的最优压缩矩阵；2.针对无线传感器网络中传感器节点能量受限这一特点,我们分别针对标量信源和向量信源两种情况,研究了网络节点的功耗分配策略。基于传感器和融合中心存在同步误差这一现象,我们提出了对应的功耗分配策略,并得到了最优分配系数的闭式解。同时,我们也做了大量的仿真来验证该算法的性能,结果表明该算法优于平均能量分配的算法；3.现有的大部分关于数据融合的算法都是在基于信道状态信息在融合中心处完全已知的假设下进行研究的,但是在实际应用中这一假设很难达到,估计过程就意味着信道状态信息的不确定。针对这一现象,我们重点研究信道估计误差对估计融合的影响,以及训练数据和传输数据之间的能量最优分配,并对线性最小均方误差(Linea minimum mean square error, LMMSE)和最优的加权最小二乘(Optimal Weighted Least squares, OWLS)两种融合算法进行了大量的分析和仿真。
[Abstract]:Wireless Sensor Networks (WSNs) as an interdisciplinary subject, which combines sensor technology, wireless communication and computer network, has attracted wide attention in recent years and has become a research hotspot at home and abroad. It has also been widely used in many military and non-military fields, such as positioning and tracking, remote control, military reconnaissance, etc. However, due to some shortcomings of the sensor itself, such as energy, bandwidth and data processing capability, In order to solve the problem of bandwidth and energy constraints and improve the efficiency of information collection, data fusion data fusion technology plays a very important role. There are still many challenges in applying data fusion to practice: the power allocation problem of sensor system, the uncertainty of channel information and so on. It makes the system achieve the estimation performance of centralized fusion while reducing the amount of transmission, and studies the optimal power allocation of the system under the condition of synchronization error. And the influence of channel estimation error on the performance of the system. The specific work includes the following contents: 1. For the measurement of noise between sensors, We propose a distributed fusion algorithm based on linear transformation. The assumption of correlated noise is more in line with the actual situation. We propose a lossless criterion of linear transformation based on sensor node measurements for linear fusion systems, and give the necessary conditions to achieve centralized fusion performance. We also give the optimal compression matrix when the equivalent measurement matrix is row full rank and column full rank. In view of the energy limitation of sensor nodes in wireless sensor networks, we consider the scalar source and vector source, respectively. In this paper, the power allocation strategy of network nodes is studied. Based on the synchronization error between sensor and fusion center, we propose a corresponding power allocation strategy and obtain the closed solution of the optimal allocation coefficient. We also do a lot of simulations to verify the performance of the algorithm. The results show that this algorithm is superior to the average energy allocation algorithm. 3. Most of the existing algorithms for data fusion are based on the assumption that the channel state information is completely known at the fusion center. However, this assumption is difficult to achieve in practical application, and the estimation process means that the channel state information is uncertain. In view of this phenomenon, we focus on the influence of channel estimation error on estimation fusion. And the optimal allocation of energy between the training data and the transmitted data. Two fusion algorithms, Linea minimum mean square error (LMMSE) and optimal weighted least squares optimal Weighted Least squares (owl LSs), are analyzed and simulated.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5

【相似文献】