无线传感器网络节点的三维定位算法研究

【作者】 李川；

【导师】 胡光岷；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2010， 硕士

【摘要】 在无线传感器网络应用中,为了关联监测事件和其发生位置,需要进行节点定位。现有无线传感器网络节点定位算法大都假设所有节点在一个二维平面上,基本思路是利用距离、角度等关系建立多个方程组或一个大型方程组,求解得到节点位置坐标,这样的算法我们称为二维定位算法。在实际应用中,无线传感器网络节点应在一个三维空间中,节点的定位是在三维空间中定位,也就是说我们应该研究三维定位算法。在三维定位算法中,由于每一个节点增加了一个未知数且测量值精度受各种因素的影响,导致方程组解的唯一性和稳定性受到影响,因此目前少见使用的三维定位算法。本文的主要思想是通过引入已知的地形信息作为约束,改善三维定位算法的准确性,由于电子地形图已相当普及,所以此思路有其实际意义。我们首先根据特点不同将定位过程分为两个阶段:一、网络初始化后的全局定位阶段,这一阶段已知少量锚节点信息,需对大量未知节点进行初始定位;二、在网络稳定后节点的移动重定位阶段,这一阶段中已知大部分参考节点信息,仅需对个别移动节点进行重定位。1、在初始的全局定位阶段,本文改进了基于锚节点推广的无线传感器网络节点定位算法。首先设计了三维定位满足的几何关系,提出新的节点误差判断标准和参考节点选择标准。本文在方程组求解过程中引入地形信息进行优化,并提出了两种算法改进策略减少定位误差。此方法的优点是仅需要少量的锚节点就能达到节点的全局定位。2、在节点移动重定位阶段,本文推广了APIT算法并设计相关协议完成定位。首先设计了利用APIT进行三维移动重定位协议,此协议规定了具体的节点行为、记录表项设计等,并为后续计算收集必要信息。然后对APIT算法进行三维推广。在具体的节点定位时,采用了结合地形信息的最优化方法,兼顾定位效率及准确率。这一方法的优点是定位简单,能较快完成定位。文中提出的算法均在Windows平台下利用Matlab进行了仿真,从仿真结果可以看出,在两种定位情况下,所提出的无线传感器网络三维定位方法都取得了令人满意的结果,验证了算法的有效性。本文所提出的算法因为更有针对性,从而在地形复杂或者大范围节点定位情况下能够较好适用。更多还原

【Abstract】 With the development of sensor technology, detection technology, and the wireless network technologies, Wireless Sensor Network technology is playing an increasingly important role in scientific researches. The positioning algorithms of WSN are needed in order to connect the detection information of sensors to the real locations. The current positioning algorithms of wireless sensor network are mostly under the two-dimensional assumption, and they can perform well in small regions or flat areas. When under the three-dimensional assumptions, errors introduced by many factors such as communication barriers and higher dimensions will reduce the accuracy of positioning algorithms. Therefore, it is a very practical and significant project to promote the mature two-dimensional positioning algorithms to three-dimensional fields.According to the different targets of positioning, this issue of wireless sensor network can be divided into two stages. The first stage is to identify the entire network after the initial layout of sensors. The latter stage is to locate individual node after the entire network has been identified. According to the different objectives of two phases, this paper uses two customized WSN (Wireless Sensor Network) three-dimensional location methods for the two conditions.In the initial phase of the global positioning, we improved a two-dimensional algorithm based on anchor promotions. The advantage of the method is that it only requires a small amount of anchors to locate all unknown nodes in the network. Its drawback is when the anchor nodes are too rare and their distribution is relatively concentrated, the errors will increases rapidly away from the region of anchors. In extended to three-dimensional process, this paper finds the geometric relationship between nodes firstly and introduces new reference node selection criterion and new targets of optimization. As the electronic map information has been very popular at the present, so this paper introduces the map information in the optimal design to reduce the positioning errors. In addition, we propose two improved strategies to find the right location: dynamic optimization of threshold and improved anchor node selection strategy.If some individual nodes need to move after the completion of global positioning, it is necessary to design new strategy to relocate them. This paper promotes the APIT algorithm to complete the positioning under this case. The advantage of this approach is that it is fast and simple. This paper designs a protocol using the three-dimensional positioning APIT algorithm firstly, and this protocol specifies the actions of nodes and the information collected for the computing stage. Then we introduce the three-dimensional extension of APIT algorithm. First extend the geometric relationship between the regions and nodes to three-dimensional, and then introduces the map information to its optimization stages to make the method both efficiency and accuracy.Both methods above are simulated by Matlab in windows platform. The simulating results show that both algorithm and protocol can provide considerable location performance under different conditions. The proposed algorithm is more close to the actual situation because of the three-dimensional assumption, which can be applied in complex terrain or under a wide range of Wireless Sensor Networks.更多还原