基于WSN的三维定位系统与算法研究

发布时间：2018-05-20 07:57

本文选题：无线传感器网络 + 三维定位　；参考：《东北大学》2014年硕士论文

【摘要】：节点定位是无线传感器网络的关键技术和应用基础,并且定位技术已经从二维空间逐步过渡到三维空间。实际定位应用中,二维平面定位系统一般不能满足要求,必须扩展到三维空间中,其中的重点研究方向为室内三维空间的定位。在部署室内三维定位系统的过程中,往往也会遇到一些问题,如信标节点的自定位问题,解决问题的算法具有很强的实用性。三维空间有其独有的特点,算法将会比二维平面定位更加复杂,更适于三维空间的算法仍然有巨大的发展空间,值得去进行研究探索。因此,本文主要研究三维定位系统节点间相对位置关系的特点,并基于其特点提出了一些改进的算法。本文首先阐述了无线传感器网络的基本概念、结构、特点,对无线传感器网络定位技术进行了分析。然后介绍了适合三维环境的定位算法。并对Cricket 3D定位系统的原理和软硬件结构进行了简单的介绍。根据Cricket节点的特点,对信标节点的布置方式进行了研究,最后布置了三维定位系统。然后分析信标节点位置估计的线性自定位算法的应用环境与特点,并提出一种基于加权最小二乘改进的线性自定位算法,继承了线性自定位不依靠被测节点的位置坐标,只需要增加一定测量次数,就可以获得信标节点间的相对位置坐标的特点。并依据不同的几何分布对误差传递影响的不同提出了加权最小二乘估计,进一步提升了估计精度,并进行了仿真实验对提出算法性能进行了分析。其次研究了基于无线传感器网络的三维定位系统中信标节点的位置坐标优化的问题。首先对移动机器人和定位系统进行建模,提出了提高机器人定位精度的系统状态预测方程,实现了信标节点的位置校准优化。然后对比分析了EKF, HF, STF几种滤波算法性能,HF对于系统模型的不确定性有更好的适应能力,STF在系统存在较大误差扰动的情况下,对真实状态有着更好的跟踪特性,并通过仿真数据进行了验证。最后分析了基于Cricket定位系统的三维定位算法的特点,提出了三个信标节点进行三维定位的问题,并研究了它的两种解决算法：高斯牛顿迭代法和基于Cayley-Menger行列式的定位算法。高斯牛顿迭代法精度高,效率低；基于Cayley-Menger行列式的定位算法精度低,效率高。然后在三维定位系统中引入了GDOP概念,并进一步提出了更适合本系统的easy-GDOP概念,并分析了二者在信标节点选择上的应用。之后研究了GDOP和easy-GDOP值做加权数据融合各自的精度和特点,easy-GDOP加权数据融合有着更高的计算效率。
[Abstract]:Node location is the key technology and application basis of wireless sensor networks, and the localization technology has been gradually transferred from two-dimensional space to three-dimensional space. In practical positioning applications, two-dimensional planar positioning system can not meet the requirements, and must be extended to three-dimensional space, in which the focus of research is indoor three-dimensional space positioning. In the process of deploying indoor 3D positioning system, some problems are often encountered, such as the self-localization of beacon nodes, and the algorithm to solve the problem is very practical. Three-dimensional space has its unique characteristics, the algorithm will be more complex than two-dimensional plane positioning, the algorithm is more suitable for three-dimensional space still has a huge space for development, it is worth to study and explore. Therefore, this paper mainly studies the characteristics of the relative position relationship between nodes in 3D positioning system, and puts forward some improved algorithms based on its characteristics. In this paper, the basic concept, structure and characteristics of wireless sensor network (WSN) are introduced, and the location technology of WSN is analyzed. Then the location algorithm suitable for three-dimensional environment is introduced. The principle, hardware and software structure of Cricket 3D positioning system are briefly introduced. According to the characteristics of Cricket nodes, the layout of beacon nodes is studied. Finally, a three-dimensional positioning system is arranged. Then, the application environment and characteristics of the linear self-location algorithm for beacon node location estimation are analyzed, and an improved linear self-location algorithm based on weighted least squares is proposed, which inherits the position coordinates of the node that is not dependent on the measured node. The characteristics of relative position coordinates between beacon nodes can be obtained by adding a certain number of measurements. The weighted least square estimation is proposed according to the influence of different geometric distribution on error transfer. The estimation accuracy is further improved and the performance of the proposed algorithm is analyzed by simulation experiments. Secondly, the optimization of position coordinates of CITIC nodes in three-dimensional positioning system based on wireless sensor network is studied. Firstly, the mobile robot and the positioning system are modeled, and the system state prediction equation is proposed to improve the positioning accuracy of the robot, and the position calibration optimization of the beacon node is realized. Then, the performance of several filtering algorithms is compared and analyzed. The EKF, HF, STF has better adaptability to the uncertainty of the system model, and it has better tracking performance to the real state when the system has large error disturbance. It is verified by simulation data. Finally, the characteristics of 3D localization algorithm based on Cricket positioning system are analyzed, and the problem of three beacon nodes for 3D localization is put forward, and its two algorithms: Gao Si Newton iterative method and Cayley-Menger determinant based localization algorithm are studied. Gao Si Newton iteration method has high accuracy and low efficiency, and the location algorithm based on Cayley-Menger determinant has low accuracy and high efficiency. Then the concept of GDOP is introduced into the 3D positioning system, and the concept of easy-GDOP, which is more suitable for the system, is put forward, and the application of the two in the selection of beacon nodes is analyzed. Then, the accuracy and characteristics of weighted data fusion of GDOP and easy-GDOP are studied. The easy-GDOP weighted data fusion has higher computational efficiency.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5;TP212.9

【相似文献】