基于模糊聚类的蓝牙4.0室内指纹定位方法研究

发布时间：2018-05-01 22:05

本文选题：室内定位 + 位置指纹　；参考：《兰州交通大学》2017年硕士论文

【摘要】：室内定位技术是指在室内环境下,通过使用基站定位,惯性辅助定位等多种定位技术一起形成一套适用于室内的位置定位方法,实现室内环境下的位置感知。随着无线通信技术的快速发展,人们在日常生活中对于室内基于位置信息服务的需求越来越强烈,而高精度的室内定位技术是获取位置信息的关键,这使得高精度的室内定位技术成为近年来的热点课题。由于蓝牙4.0具有功耗低、成本低、便于部署的特点,本文使用其作为无线技术基础,结合有代表性的室内位置指纹定位技术进行研究和优化。室内位置指纹定位技术主要有两个阶段:信息库建立阶段和实时定位阶段。为了提高算法性能,传统优化方法先对信息库的位置信息进行聚类,然后在定位阶段使用K近邻(KNN)算法进行匹配,保证定位精度的同时减少算法计算量。本文针对传统基于聚类的位置指纹定位方法中使用的模糊C均值(FCM)算法存在的缺陷,如对初始值敏感,易陷入局部最优解等,引入了粒子群优化算法,先求解FCM算法的初始值,之后进行FCM聚类,这样避免了算法陷入局部最优解,提高了算法的定位精度和鲁棒性。同时,针对信息库建立阶段的采样成本问题,采用重心拉格朗日插值算法进行优化,在保证算法精度的同时降低采样成本;在实时定位阶段,采用模糊决策的方法,避免了KNN算法造成的误差和计算冗余。最后,本文通过Matlab平台进行了优化算法的总体设计及实验。实验对比了优化前后定位算法的性能,包括信息库建立阶段中优化前后的插值算法对定位性能的影响,优化前后聚类算法性能对比,优化前后定位算法的性能对比。根据本文实验结果分析,改进后的位置指纹定位方法在定位准确度和定位实时性上都有了较好的改善。增加插值算法后,也能在较好地保证精度的同时显著减少采样成本。
[Abstract]:Indoor positioning technology refers to the indoor environment, through the use of base station positioning, inertial auxiliary positioning and other positioning techniques together to form a set of indoor location methods, to achieve indoor location perception. With the rapid development of wireless communication technology, people need more and more indoor location-based information service in daily life, and high-precision indoor positioning technology is the key to obtain location information. This makes high-precision indoor positioning technology become a hot topic in recent years. Because Bluetooth 4.0 has the characteristics of low power consumption, low cost and easy to deploy, this paper uses Bluetooth 4.0 as the wireless technology foundation, combined with the representative indoor location fingerprint location technology to study and optimize. There are two main phases of indoor location fingerprint location: information base establishment and real-time location. In order to improve the performance of the algorithm, the traditional optimization methods first cluster the location information of the information base, and then use the K-nearest neighbor KNN algorithm to match in the localization stage to ensure the accuracy of the location and reduce the computational complexity of the algorithm. Aiming at the shortcomings of the fuzzy C-means FCM algorithm used in the traditional location fingerprint location method based on clustering, such as being sensitive to the initial value and easy to fall into the local optimal solution, the particle swarm optimization algorithm is introduced to solve the initial value of the FCM algorithm. Then FCM clustering is carried out, which avoids the algorithm falling into local optimal solution, and improves the location accuracy and robustness of the algorithm. At the same time, aiming at the cost of sampling in the stage of information base establishment, the Lagrange interpolation algorithm of gravity center is used to optimize the algorithm to ensure the accuracy of the algorithm and to reduce the sampling cost, and in the real-time positioning stage, the fuzzy decision method is adopted. The error caused by KNN algorithm and computational redundancy are avoided. Finally, the overall design and experiment of the optimization algorithm are carried out on the Matlab platform. The performance of the pre-and post-optimization localization algorithm is compared, including the influence of the pre-and post-optimization interpolation algorithm on the location performance, the performance comparison of the pre-and post-optimization clustering algorithm, and the performance comparison of the pre-and post-optimization localization algorithm. According to the experimental results of this paper, the improved location fingerprint localization method has better accuracy and real time. After adding the interpolation algorithm, the sampling cost can be significantly reduced while the precision is better guaranteed.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN925

【参考文献】