基于无线传感器网络的室内定位技术研究

发布时间：2018-03-02 11:02

本文关键词： 无线传感器网络室内定位最小二乘算法质心算法 DV-Hop定位算法　出处：《江南大学》2017年硕士论文　论文类型：学位论文

【摘要】：近年来,随着无线传感器网络技术的不断发展以及人们对于室内定位的迫切需求,传感器网络的易于部署、可扩展性高、成本低等优点,使得基于无线传感器网络的室内定位技术受到广泛关注。然而,由于室内环境复杂多变造成信号传播反射、散射和遮蔽等影响,传统的传感器定位技术在室内难以获得精准而有效的位置信息。目前基于无线传感器网络的室内定位技术逐渐由理论迈向实际应用。但仍面临许多亟待解决的问题,其中包括如何解决定位过于依赖锚节点的密度,如何有效地利用测距技术,如何优化定位算法中的非线性计算问题,如何平衡定位精度、计算复杂度和定位稳定性之间的关系。基于上述问题,本文的研究内容主要包括以下几个方面:(1)针对最小二乘算法定位过程中易受到测距误差影响进行研究。分析了粒子群算法的缺陷,提出一种基于自适应差分的粒子群定位算法改进方案。该算法首先利用环境补偿拟合未知节点到锚节点的距离,求解适应值函数;然后利用改进的自适应差分算法生成新的种群,再用粒子群算法和新的变异策略进行局部搜索,与适应值比较反复迭代逐渐收敛;最后得出未知节点位置。仿真结果表明,在室内定位中,与其它改进粒子群算法相比改进后的算法在抑制误差累积与定位效果方面有了极大地改善。(2)针对质心定位算法中定位精度低的问题进行研究。分析了现有的质心定位算法存在的缺陷,提出一种基于最大似然估计加权的质心定位算法改进方案。该算法首先把锚节点与未知节点之间距离的最大似然估计值作为权值;然后在权值模型中,引进一个参数k优化未知节点周围锚节点分布;最后对未知节点的估计位置修正。仿真结果表明,基于最大似然估计的加权质心算法具有定位精度高和成本低的特点,优于基于距离倒数的质心加权和基于RSSI倒数的质心加权算法,适用于大面积的室内定位。(3)针对DV-Hop定位算法中误差累积的问题进行研究。分析了传统的DV-Hop定位中的缺陷,提出一种基于跳数和跳距修正的遗传算法优化DV-Hop定位算法改进方案。该算法首先通过锚节点之间的RSSI均值对最小跳数进行约束;然后利用锚节点之间的最小跳数改善平均跳距;最后利用改进的遗传算法优化位置估计结果。仿真结果表明,与基于跳距加权DV-Hop算法和遗传优化的DV-Hop算法相比,定位精度有明显提高。综上所述,本文针对几种室内定位算法的定位效果差的问题,致力于改进定位精度的研究并提出相应的改进方案,利用MATLAB对其仿真验证。实验结果表明,改进后的算法在定位误差方面有了一定的改善,定位效果与同类型的算法相比明显优越。
[Abstract]:In recent years, with the continuous development of wireless sensor network technology and the urgent need for indoor positioning, sensor networks are easy to deploy, high scalability, low cost and so on. The indoor localization technology based on wireless sensor network has attracted wide attention. However, because of the complex indoor environment, the effects of signal propagation, reflection, scattering and shadowing are caused by the complexity of indoor environment. Traditional sensor positioning technology is difficult to obtain accurate and effective location information in indoor. At present, the indoor location technology based on wireless sensor network is gradually moving from theory to practical application. However, there are still many problems to be solved. It includes how to solve the density of anchor nodes, how to make use of ranging technology effectively, how to optimize the nonlinear calculation problem of location algorithm, how to balance the accuracy of location, and how to solve the problem of how to solve the problem. The relationship between computational complexity and location stability. The main contents of this paper include the following aspects: 1) to study the vulnerability of the least-squares algorithm to ranging errors in the localization process, and to analyze the defects of the particle swarm optimization algorithm (PSO). An improved Particle Swarm Optimization (PSO) algorithm based on adaptive difference is proposed. Firstly, the distance from unknown node to anchor node is fitted by environment compensation, and the fitness function is solved. Then the improved adaptive difference algorithm is used to generate the new population, then the particle swarm optimization algorithm and the new mutation strategy are used to carry out local search. Compared with the fitness value, iterative iteration gradually converges. Finally, the unknown node position is obtained. The simulation results show that, In indoor positioning, Compared with other improved particle swarm optimization algorithm, the improved algorithm has greatly improved the error accumulation and localization effect. (2) the problem of low positioning accuracy in centroid localization algorithm is studied. The existing centroid is analyzed. The shortcomings of the localization algorithm, An improved centroid localization algorithm based on weighted maximum likelihood estimation (MLE) is proposed, in which the maximum likelihood estimation of the distance between anchor node and unknown node is taken as the weight, and then in the weight model, A parameter k is introduced to optimize the distribution of anchor nodes around unknown nodes. Finally, the estimated location of unknown nodes is modified. The simulation results show that the weighted centroid algorithm based on maximum likelihood estimation has the characteristics of high positioning accuracy and low cost. Better than the centroid weighting algorithm based on the reciprocal distance and the centroid weighting algorithm based on the RSSI reciprocal, it is suitable for large area indoor positioning. (3) the problem of error accumulation in the DV-Hop location algorithm is studied, and the defects in the traditional DV-Hop location are analyzed. A genetic algorithm based on number of hops and modified hops is proposed to optimize the DV-Hop localization algorithm. Firstly, the minimum hops are constrained by the RSSI mean between anchor nodes, and then the average hops are improved by using the minimum hops between anchor nodes. Finally, the improved genetic algorithm is used to optimize the location estimation results. The simulation results show that compared with the hopping weighted DV-Hop algorithm and the genetic optimization DV-Hop algorithm, the positioning accuracy is obviously improved. Aiming at the problem of poor localization effect of several indoor positioning algorithms, this paper is devoted to the research of improving the positioning accuracy and puts forward the corresponding improvement scheme, which is verified by MATLAB. The experimental results show that, The improved algorithm has a certain improvement in localization error, and the localization effect is obviously superior to that of the same kind of algorithm.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5;TP212.9

【引证文献】