基于HBase的面向语义单元的室内移动对象索引

发布时间：2018-04-19 17:15

本文选题：室内 + 移动对象　；参考：《地球信息科学学报》2017年03期

【摘要】：随着室内定位技术的广泛应用,传感器记录了大量室内移动对象的位置数据,而索引技术作为移动对象数据分析的基础工作也得到越来越多的研究。已有索引技术多是针对室外空间的移动对象,不能支持室内移动对象数据的三维立体性、轨迹的复杂性、随机性等特点,这些索引技术也仅仅关注了移动对象的位置信息,忽略了语义信息,不能有效地支持室内移动对象的管理和分析,并且当面对海量的移动对象数据时,这些架构在传统关系型数据库上的索引都存在性能瓶颈问题。因此,本文提出了面向语义单元的移动对象表达模型,利用语义单元将室内移动对象的位置语义化,设计了SCo II(Semantic Cell Oriented Indoor moving objects Index)索引结构对室内移动对象的历史数据进行索引,能够有效支持语义粒度上的时空范围查询、移动对象语义轨迹查询。索引基于HBase实现,能够适应大规模的并发更新与查询,具有良好的规模扩展性,规避了大数据给传统数据库带来的性能瓶颈问题,实验证明其具有良好的更新和查询性能。该索引的实现方便了基于语义的室内移动对象分析和数据挖掘工作,为今后的分析工作奠定了基础。
[Abstract]:With the wide application of indoor positioning technology, sensor records a large number of indoor moving object position data, and indexing technology as the basic work of moving object data analysis has been more and more studied.Most of the existing indexing techniques are aimed at moving objects in outdoor space, which can not support the three-dimensional, complex and random characteristics of indoor moving object data. These indexing techniques only focus on the location information of moving objects.The semantic information is ignored, and the management and analysis of indoor moving objects can not be effectively supported. In the case of massive moving object data, there are performance bottlenecks in the indexes of these architectures on traditional relational databases.Therefore, a semantic unit oriented moving object representation model is proposed in this paper. The SCo II(Semantic Cell Oriented Indoor moving objects Index (SCo II(Semantic Cell Oriented Indoor moving objects Index) index structure is designed to index the historical data of the moving object.It can effectively support spatio-temporal query on semantic granularity and semantic track query of moving object.The index is implemented based on HBase and can adapt to large-scale concurrent update and query. It has good scale expansibility and avoids the performance bottleneck problem brought by big data to the traditional database. Experiments show that the index has good update and query performance.The implementation of this index facilitates the analysis and data mining of indoor moving objects based on semantics and lays the foundation for the future analysis work.
【作者单位】：中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室;中国科学院大学;
【基金】：国家自然科学基金项目(41590845) 山西省-中国科学院科技合作项目(20141011001)
【分类号】：TP311.13

【相似文献】