云计算环境下图书馆位置感知服务研究

Study on Library Location Awareness Service under Cloud Computing Environment

针对采用基于WLAN的位置指纹算法的图书馆位置感知服务器在处理大规模定位数据时计算时间太长,结合云计算技术的快速发展,提出一种基于MapReduce架构的并行朴素贝叶斯概率计算方法。此方法通过MapReduce框架对指纹数据库可快速计算出定位目标在每个记录的后验概率值,以后验概率为权重来确定位置坐标,使定位在准确性和时间响应性方面获得更好的性能,以提升图书馆位置感知服务质量。文章将并行朴素贝叶斯概率算法进行图书馆室内定位实验,并将其与串行朴素贝叶斯概率算法对比测试。实验结果表明,改进后的定位算法有效减短了定位过程的计算时间,而且还能保证定位精度,完全满足定位服务需求。

Combining with the rapidly-developing cloud computing technology,this paper proposes a parallel naive Bayesian probability calculation method based on MapReduce architecture,so as to solve the problem that library location awareness server using WLAN-based location fingerprinting algorithm takes too long when processing large-scale location data.This method can quickly calculate fingerprint database and work out the posterior probability of located targets at each record through MapReduce architecture,and then determine the location coordinates of targets with the posterior probability obtained as weight,thus improving the performance of library location awareness server in terms of location accuracy and time responsiveness and perfecting the quality of library location awareness services.Besides,in this paper,library indoor location experiment using the proposed parallel naive probability calculation method is conducted,and comparison experiment is made between the proposed method and serial naive Bayesian probability calculation method.Experiment results show that the improved location calculation method can not only effectively shorten the computing time of location process,but also ensure location accuracy so as to fully meet location service requirements.