基于 HBase 的面向语义单元的室内移动对象索引

doi:10.3724/SP.J.1047.2017.00307

摘要/Abstract

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SCoII （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。

关键词: 室内, 移动对象, 索引, 语义, HBase

Abstract:

With the development of indoor positioning technique, more and more position data of indoor moving objects are recorded by sensors. As the basic work of moving objects database, index technique has become a research hot-spot. Majority of existing moving objects index are for outdoor moving objects which are not suitable for indoor environment. Also, they only build index on geography coordinates of moving objects, lack of supporting of semantic information which can offer effective support for management and analysis of indoor moving objects. There will be a performance bottleneck when massive data are ingested and frequent querying are asked when implemented on traditional relational database. In this paper, we built a grid of indoor floor environment and create a map relation from grid to semantic cell. Then, we utilized this map to semanticize indoor moving objects’ location if it was contained in a semantic cell. After this work, we built an index called SCoII （Semantic Cell Oriented Indoor moving objects Index）. SCoII can answer not only semantic spatio-temporal range query but also indoor moving object’s semantic trajectory query, which can support for semantic-based analysis of indoor moving objects. SCoII is implemented on HBase, so it also avoided the performance degradation of traditional relational database when encounting massive data and have good performance of updating and querying without bottleneck. Experimental results also showed that it can be adapt to big data. Supporting for semantic information of indoor moving object is the most important feature of SCoII. More data mining jobs can be done on indoor moving object’s semantic location and semantic trajectory such as the simple example given out at the end. Management and analysis based on semantic of indoor moving objects will be convenient on SCoII, which lays a foundation of analysis work in the future.

Key words: indoor, moving objects, index, semantic, HBase

张得群, 谢传节, 裴韬. 基于 HBase 的面向语义单元的室内移动对象索引[J]. 地球信息科学学报, 2017, 19(3): 307-316.DOI:10.3724/SP.J.1047.2017.00307

ZHANG Dequn,XIE Chuanjie,PEI Tao. Semantic Cell Oriented Indoor Moving Objects Index based on HBase[J]. Journal of Geo-information Science, 2017, 19(3): 307-316.DOI:10.3724/SP.J.1047.2017.00307

图/表 10

表1

图1

表3

图2

图3

图4

表2

图5

图6

图7

参考文献 23

[1]	王倩. 室内移动对象轨迹相似性度量与应用[D].合肥:中国科学技术大学,2015.
	[ Wang Q.Similarity measurement and application of indoor moving-object trajectories[D]. Hefei: University of Science and Technology of China, 2015. ]
[2]	Ying J C, Lu H C, Lee W C, et al.Mining user similarity from semantic trajectories. Proceedings of the 2nd ACM SIGSPATIAL international workshop on location based social networks[C]. ACM, 2010:19-26.
[3]	Hwang J R, Kang H Y, Li K J.Spatio-temporal similarity analysis between trajectories on road networks[M]. Springer Berlin Heidelberg, 2005:280-289.
[4]	廖律超,蒋新华,邹复民,等.一种支持轨迹大数据潜在语义相关性挖掘的谱聚类方法[J].电子学报,2015,43(5):956-964. doi: 10.3969/j.issn.0372-2112.2015.05.019
	[ Liao L C, Jiang X H, Zou F M, et al.A spectral clustering method for big trajectory data mining with latent semantic correlation[J]. Acta Electronic Sinica, 2015,43(5):956-964. ] doi: 10.3969/j.issn.0372-2112.2015.05.019
[5]	齐凌艳,陈荣国,温馨.基于语义轨迹停留点的位置服务匹配与应用研究[J].地球信息科学学报,2014,16(5):720-726. doi: 10.3724/SP.J.1047.2014.00720
	[ Qi L Y, Chen R G, Wen X.Research on the LBS matching based on stay point of the semantic trajectory[J]. Journal of Geo-Information Science, 2014,16(5):720-726. ] doi: 10.3724/SP.J.1047.2014.00720
[6]	方颖. 移动对象数据库中移动对象索引方法研究[D].武汉:武汉大学,2010.
	[ Fang Y.Research on moving objects indexing methods in moving objects databases[D]. Wuhan: Wuhan University, 2010. ]
[7]	贲婷婷,秦小麟,王丽.基于语义和访问权限的室内移动对象索引[J].计算机科学,2015,42(3):178-184. doi: 10.11896/j.issn.1002-137X.2015.3.037
	[ Ben T T, Qin X L, Wang L.Index of indoor moving objects based on semantics and access permission[J]. Computer Science, 2015,42(3):178-184. ] doi: 10.11896/j.issn.1002-137X.2015.3.037
[8]	杨彬. 室内移动对象的数据管理[D].上海:复旦大学,2010.
	[ Yang B.Data management of indoor moving objects[D]. Shanghai: Fudan University, 2010. ]
[9]	Lu H, Yang B, Jensen C S.Spatio-temporal joins on symbolic indoor tracking data. Data Engineering (ICDE), 2011 IEEE 27th International Conference on[C]. IEEE, 2011:816-827.
[10]	汪娜. 面向室内空间的时空数据管理关键技术研究[D].合肥:中国科学技术大学,2014.
	[ Wang N.Research on key techniques of spatio-temporal data management for indoor space[D]. Hefei: University of Science and Technology of China, 2014. ]
[11]	冯晓普. HBase存储的研究与应用[D].北京:北京邮电大学,2014.
	[ Feng X P.Research and application of the storage of HBase[D]. Beijing: Beijing University of Posts and Telecommunications, 2014. ]
[12]	Zhang N, Zheng G, Chen H, et al.HBasespatial: A scalable spatial data storage based on HBase. 2014 IEEE 13th international conference on trust, security and privacy in computing and communications[C]. IEEE, 2014:644-651.
[13]	Jensen C S, Lu H, Yang B.Indexing the trajectories of moving objects in symbolic indoor space[M]. Advances in Spatial and Temporal Databases. Springer. 2009:208-227
[14]	甘早斌,袁永光,赵贻竹,等.基于DR-tree的室内移动对象索引研究[J].计算机科学,2012,39(10):177-181. doi: 10.3969/j.issn.1002-137X.2012.10.039
	[ Gan Z B, Yuan Y G, Zhao Y Z, et al.Indoor moving objects index research based on DR-tree[J]. Computer Science, 2012,39(10):177-181. ] doi: 10.3969/j.issn.1002-137X.2012.10.039
[15]	金培权,汪娜,张晓翔,等.面向室内空间的移动对象数据管理[J].计算机学报,2015,38(9):1777-1795. doi: 10.11897/SP.J.1016.2015.01777
	[ Jin P Q, Wang N, Zhang X X, et al.Moving object data management for indoor spaces[J]. Chinese Journal of Computers, 2015,38(9):1777-1795. ] doi: 10.11897/SP.J.1016.2015.01777
[16]	Shin S, Kim G, Bae H.Adaptive cell-based index for moving objects in indoor[J]. KSII Transactions on Internet and Information Systems (TIIS), 2012,6(7):1815-1830. doi: 10.3837/tiis.2012.07.006
[17]	贲婷婷,秦小麟,许建秋.支持多种查询的室内移动对象索引[J].计算机研究与发展,2015,52(9):2002-2013. doi: 10.7544/issn1000-1239.2015.20131230
	[ Ben T T, Qin X L, Xu J Q.Index of indoor moving objects for multiple queries[J]. Journal of Computer Research and Development, 2015,52(9):2002-2013. ] doi: 10.7544/issn1000-1239.2015.20131230
[18]	Li S, Hu S, Ganti R, et al.Pyro: A spatial-temporal big-data storage system. 2015 USENIX Annual Technical Conference (USENIX ATC 15)[C]. 2015:97-109.
[19]	Nishimura S, Das S, Agrawal D, et al.MD-HBase: design and implementation of an elastic data infrastructure for cloud-scale location services[J]. Distributed and Parallel Databases, 2013,31(2):289-319. doi: 10.1007/s10619-012-7109-z
[20]	Nishimura S, Das S, Agrawal D, et al.MD-HBase: a scalable multi-dimensional data infrastructure for location aware services. Mobile Data Management (MDM), 2011 12th IEEE International Conference[C]. IEEE, 2011:7-16.
[21]	Hughes J N, Annex A, Eichelberger C N, et al.GeoMesa: A distributed architecture for spatio-temporal fusion. SPIE Defense+ Security[C]. International Society for Optics and Photonics, 2015.
[22]	Whitman R T, Park M B, Ambrose S M, et al.Spatial indexing and analytics on Hadoop. ACM sigspatial international conference on advances in geographic information systems[C]. ACM, 2014:73-82.
[23]	Chen X, Zhang C, Ge B, et al.Spatio-temporal queries in HBase[C]. 2015 IEEE International Conference on Big Data, 2015:1929-1937.

索引分类	索引名称	索引结构	支持的查询种类
基于传统数据库的移动对象索引	RTR-TP²R-tree	R-tree	室内时空范围查询、轨迹查询
	DR-tree	R-tree	室内对象轨迹查询
	ACII	先 R-tree后 Hash	全时态移动对象查询
	MQII	Hash后使用链表指针	对象查询、范围查询
基于分布式数据库的移动对象索引	MD-HBase	KD-tree/Quad-tree + Z-ordering	时空范围查询
	Pyro	Moore	时空范围查询
	GeoMesa	Z-ordering	对象查询、时空范围查询
	STEHIX	Hilbert + Quad-tree + 分段索引	时空范围查询,kNN

Cell名称	进入时间	离开时间	停留时长/s
Rouge Diamant	12:21:24	12:21:26	2
过道	12:21:26	12:21:28	2
Rouge Diamant	12:21:28	12:21:31	3
Bcuthentique	12:21:31	12:22:29	58
FIVE PLUS+	12:23:14	12:24:01	47
TRENDIANO	12:27:51	12:50:04	1333
Levi's ladies	12:50:04	13:04:53	889
Jack jones	13:27:16	13:28:16	60
TEENIE&WEENIE	13:28:58	13:30:56	118
TEENIE&WEENIE	13:41:22	17:43:14	14 512
H&M	17:43:14	17:57:07	833

字段	含义
x	X坐标值
y	Y坐标值
floor	定位点所在的楼层,20040表示F4,10020表示B2
time	数据采集的时间,精确到秒
mac	定位对象设备的物理地址