基于TDOA的无线传感器网络定位算法研究

发布时间：2018-03-08 21:05

本文选题：无线传感器网络　切入点：节点定位　出处：《南京航空航天大学》2014年硕士论文　论文类型：学位论文

【摘要】：作为无线传感器网络关键技术之一,节点定位是众多应用的基础与核心技术,在无线传感器网络的监测活动中的作用至关重要。基于TDOA测距的节点定位算法目前较为广泛使用,此类算法有两个主要误差源:非视距导致的TDOA测距误差以及非线性方程解的误差。本课题研究机场飞行区内特种车辆的定位问题。飞行区地面部分包括跑道、滑行道、停机坪等,由于大量飞机、特种车辆的存在,此区域NLOS误差严重且误差模型无法确定。在基于TDOA的无线定位算法中,为了有效抑制NLOS造成的TDOA测距误差,本文针对卡尔曼滤波算法进行了改进。通过研究误差特性,针对较大NLOS误差会导致后续估计值均受严重影响引入新息阈值,而针对NLOS误差不可避免的特性引入修正系数,进而改进卡尔曼滤波迭代过程,以获得更准确的估计值。通过仿真与原始卡尔曼滤波算法以及部分改进算法进行了对比,证实了本文算法的有效性。其次,在基于TDOA的无线传感器网络定位算法中,最小二乘算法复杂度低,但抗非视距误差能力较弱。Taylor级数展开算法能够通过迭代求精有效抑制NLOS误差,但受初始估计值影响较大。本文针对两种算法缺点及机场特种车定位特性,将两种算法进行了混合。首先利用最小二乘算法获得初始估计位置,然后并取该估计域质心作为Taylor算法初始值进行迭代求精。通过仿真与最小二乘法及Taylor算法进行了对比,结果表明改进算法具有更高的定位精度。
[Abstract]:As one of the key technologies of wireless sensor networks, node localization is the foundation and core technology of many applications, play an important role in monitoring the activities of wireless sensor networks. The node localization algorithm based on TDOA is currently used widely, this kind of algorithm has two main error sources: non line of sight caused by the error of TDOA algorithm and nonlinear equations solution the problem of positioning error. This paper studies the airport flight area special vehicles. The flight area includes ground runway, taxiway, apron, due to a large number of aircraft, special vehicle, this area is serious and the error model of NLOS error can not be determined. Based on the TDOA wireless location algorithm, in order to effectively suppress the ranging error of TDOA NLOS, the Calman filter method. Through the study of the characteristics of error for large NLOS error will lead to subsequent estimates Are affected by the introduction of new information and according to the characteristics of NLOS error threshold, inevitably the correction coefficient, and then improved Calman filter iterative process, in order to get more accurate estimation values. By simulation with the original Calman filtering algorithm and some improved algorithms were compared to show the effectiveness of the algorithm. Secondly, based on the location of TDOA the wireless sensor network algorithm, least squares algorithm with low complexity, but the NLOS error mitigation capability of.Taylor series expansion algorithm can effectively restrain the NLOS errors through iterative refinement, but the initial estimate value. Aiming at the influence of two kinds of special vehicle positioning algorithm shortcomings and airport characteristics, the two algorithms are firstly mixed. Least squares algorithm to obtain the initial position estimate, and then the estimation of domain of centroid iterative refinement as the initial value of Taylor algorithm by simulation. Compared with the least square method and the Taylor algorithm, the results show that the improved algorithm has higher positioning accuracy.

【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP212.9;TN929.5

【参考文献】