带通信和测量驱动的传感器网络分布式估计算法设计

 2022-05-05 08:05

论文总字数:42202字

摘 要

作为新兴的研究热点,无线传感器网络在越来越多的领域有了更广泛的应用,涉及目标跟踪、军事、医疗、精细农业等多个方面。由于传感器网络节点本身的对于信息的处理能力有限,且处理信息时对能量的消耗巨大,因此,采用快捷高效的目标跟踪的算法对于无线传感器网络的寿命和处理能力都十分重要。由于无线传感器网络的传感器节点对于数据的处理能力相对较弱,因此采用分布式算法能够对这个问题进行合理的解决。本文从对于跟踪目标的能量消耗、跟踪精度出发,分别设计测量与通信事件触发机制以及分布式滤波算法,在实现目标跟踪的同时有效地节省了传感器节点对于能量的要求。

一方面考虑具有线性时不变动态的移动目标的跟踪问题,采用基于最小迹原则的一致性卡尔曼滤波算法,分别设计事件触发测量机制和事件触发通信机制。算法中,对于无线传感器中的传感器节点采取两种工作方式:工作或休眠,只有当移动目标进入可测量范围内时无线传感器网络中的节点才会进行测量工作。之后,只有当节点之间的信息传递满足要求时,节点之间才会进行对于移动目标信息的传递和处理。最后,结合分布式卡尔曼滤波算法,对数据进行不断的更新与融合,最终得到关于目标移动的位置信息。仿真结果说明了该算法能够合理的满足目标跟踪的要求。

另一方面考虑具有非线性动态的移动目标跟踪问题,设计基于最小迹原则的一致性扩展卡尔曼滤波算法,并分别设计了事件触发测量机制和事件触发通信机制,实现了目标跟踪精度和节点能耗的平衡。

关键词:无线传感器网络 分布式卡尔曼滤波 事件触发 目标跟踪

Abstract

As an emerging research hotspot, wireless sensor networks have been widely used in more and more fields, including target tracking, military, medical, and fine agriculture. Since the sensor network node itself has limited processing capability for information and consumes a large amount of energy when processing information, the algorithm of fast and efficient target tracking is very important for the life and processing capability of the wireless sensor network. Since the sensor nodes of the wireless sensor network have relatively weak processing ability for data, a distributed algorithm can be used to solve this problem reasonably. In this paper, starting from the energy consumption and tracking accuracy of the tracking target, combined with measurement triggering and communication, the energy requirements of the sensor node are effectively saved.

On the one hand, considering the tracking problem of moving targets with linear time-invariant dynamics, the consistent Kalman filter algorithm based on the minimum trace principle is used to design the event trigger measurement mechanism and the event trigger communication mechanism. In the algorithm, two working modes are adopted for the sensor nodes in the wireless sensor: work or sleep, and the nodes in the wireless sensor network perform measurement only when the moving target enters the measurable range. After that, only when the information transfer between the nodes meets the requirements, the transfer and processing of the mobile target information will be performed between the nodes. Finally, combined with the distributed Kalman filter algorithm, the data is continuously updated and merged, and finally the position information about the target movement is obtained. The simulation results show that the algorithm can meet the requirements of target tracking reasonably.

On the other hand, considering the moving target tracking problem with nonlinear dynamics, a consistent extended Kalman filter algorithm based on the minimum trace principle is designed. The event trigger measurement mechanism and event trigger communication mechanism are designed respectively to achieve the target tracking accuracy and the balance of consumption.

KEY WORDS: wireless sensor network, distributed kalman filter, event trigger ,target tracking

目 录

摘 要 I

Abstract II

第一章 绪论 1

1.1 无线传感器网络背景 1

1.2 无线传感器网络研究现状 2

1.2.1 无线传感器网络目标跟踪算法研究现状 2

1.2.2 无线传感器网络事件触发机制研究现状 4

1.3 无线传感器网络目标跟踪评价指标 4

1.4 研究内容与论文构成 5

第二章 预备知识 7

2.1 滤波算法 7

2.2 卡尔曼滤波 7

2.2.1 经典卡尔曼滤波算法 7

2.2.2 扩展卡尔曼滤波算法 9

2.3 测量触发机制 11

2.3.1 双圆跟踪模型 12

2.3.2 传送树跟踪模型 13

2.4 分布式通信触发机制 14

2.4.1 基本分布式事件触发一致性算法 14

2.4.2 带有最小采样粒度的事件触发一致性算法 15

2.5 小结 18

第三章 分布式事件触发卡尔曼滤波算法 19

3.1 引言 19

3.2 问题描述 19

3.3 事件触发的分布式卡尔曼滤波算法 19

3.4 仿真实验 23

3.5 小结 27

第四章 分布式事件触发扩展卡尔曼滤波算法 28

4.1 问题描述 28

4.2 事件触发分布式扩展卡尔曼滤波算法 28

4.3 小结 32

参考文献 33

致 谢 36

绪论

无线传感器网络背景

无线传感器网络(Wireless Sensor Networks,WSN)被认为是世纪最重要的技术之一[1],它是随着自动控制技术与分布式网络的发展而逐渐受到各个领域的广泛关注。由于传感器本身拥有体积小,方便隐藏,价格低廉等的优点,并且传感器本身采用的无线传输的通信方式,所以由传感器组成的无线传感器网络技术得到了很大的关注,也因此获得了很好的发展。

剩余内容已隐藏,请支付后下载全文,论文总字数:42202字

您需要先支付 80元 才能查看全部内容!立即支付

该课题毕业论文、开题报告、外文翻译、程序设计、图纸设计等资料可联系客服协助查找;