UWSN中的单步延迟无序量测数据融合算法的研究

发布时间：2018-03-03 21:41

本文选题：水下无线传感器网络　切入点：无序量测　出处：《天津工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：在水下随机或有规律部署大量低功耗、具有一定的通信能力的传感器节点,采集水下信息,利用节点的自组织能力,形成水下无线传感器网络(Underwater Wireless Sensor Network, UWSN)。UWSN为水下环境监测提供新的手段,但UWSN通信采用声波通信,不同于传统的无线传感器网络(WSN),带来了UWSN的定位技术、网络拓扑、路由协议、水声通信和数据融合等方面的问题。UWSN的数据融合目的是通过融合减少数据的传输量,降低节点因数据传输带来的能耗,对延长UWSN的生命周期具有重要的现实意义。水声信道具有传播延迟高且动态变化、通信带宽有限、多径效应严重等特点,受自身物理条件的制约,UWSN中易出现节点先发的信息后到融合中心,后发的信息先到融合中心的现象,即无序量测(Out-of-Sequence Measurement, OOSM)现象。因此,在UWSN中,针对OOSM现象,如何融合处理带OOSM数据,直接关系到融合结果的可信度,也促使数据融合理论的完善。本文从单传感器和多传感器方面研究UWSN网中的OOSM数据融合问题,主要工作如下：(1)使用OPNET对水声信道进行仿真,建立水声通信网络模型,分析网络延迟,观测无序量测现象。(2)对UWSN环境下的单传感器单步延迟无序量测融合算法进行了研究,提出基于后向预测的无序量测融合算法。在过程噪声直接离散化模型条件下,该算法能够处理量测噪声与同一时刻过程噪声相关的单传感器单步延迟无序量测数据融合问题。该算法可以保证实时性,而且滤波精度高于A1算法,仿真结果验证了算法的有效性。(3)针对UWSN环境下的多传感器单步延迟无序量测问题,提出一种基于等价量测的单步延迟无序量测融合算法。该算法采用分布式融合结构,利用一步预测的等价量测处理延迟量测,把无序量测信息融合问题转化成一种顺序量测信息融合问题,采用噪声相关的Kalman滤波算法获得局部估计,最后使用矩阵加权算法得到全局估计。该算法的滤波精度与直接更新法相比有所降低,但计算量减少。
[Abstract]:A large number of sensor nodes with low power consumption and certain communication capability are deployed randomly or regularly under water to collect underwater information and make use of the self-organizing capability of the nodes. The formation of underwater Wireless Sensor Network, UWSN).UWSN provides a new means for underwater environment monitoring, but the UWSN communication uses acoustic communication, which is different from the traditional wireless sensor network (WSNN), and brings the location technology, network topology, routing protocol of UWSN. The purpose of UWSN data fusion is to reduce the amount of data transmission and reduce the energy consumption caused by data transmission. The underwater acoustic channel has the characteristics of high propagation delay, dynamic variation, limited communication bandwidth and serious multipath effect. Under the restriction of physical conditions, the phenomenon of node first sending information to fusion center and later information to fusion center is easy to occur in UWSN, that is, out-of-order measurement Out-of-Sequence measure (OOSMN) phenomenon. Therefore, in UWSN, the phenomenon of OOSM is discussed. How to fuse and process the data with OOSM is directly related to the credibility of the fusion results and the perfection of the theory of data fusion. In this paper, the problem of OOSM data fusion in UWSN nets is studied from the single sensor and multi-sensor aspects. The main work is as follows: (1) the underwater acoustic channel is simulated with OPNET, the network model of underwater acoustic communication is established, the network delay is analyzed, and the disordered measurement phenomenon is observed. An unordered measurement fusion algorithm based on backward prediction is proposed. Under the condition of direct discretization of process noise, The algorithm can deal with the data fusion problem of single sensor delay unordered measurement which is related to the measurement noise and the process noise at the same time. The algorithm can ensure the real-time performance and the filtering accuracy is higher than that of A1 algorithm. Simulation results show that the algorithm is effective. Aiming at the multi-sensor single-step delay disorder measurement problem in UWSN environment, a one-step delay disorder measurement fusion algorithm based on equivalent measurement is proposed. The algorithm adopts a distributed fusion structure. By using one-step prediction equivalent measure processing delay measurement, the unordered measurement information fusion problem is transformed into a sequential measurement information fusion problem, and the local estimation is obtained by using the noise-dependent Kalman filtering algorithm. Finally, the global estimation is obtained by using the matrix weighting algorithm. The filtering accuracy of the algorithm is lower than that of the direct updating method, but the computational complexity is reduced.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP212.9;TN929.3

【相似文献】