Research On Massive Data Storage And Query Techniques For Crowdsensing-based Internet Of Vehicles

As a typical application of Internet of Things(IoT)in transportation,Internet of Vehicles(IoV)exploits technologies of many fields like intelligent transportation and mobile Internet to construct an open and integrated network system of "people-car-road"integration.In this way,it realizes the monitoring and management of traffic conditions,resulting in improving traffic efficiency and safety.Crowdsensing-based IoV has solved the problem that how to collect a large scale of traffic data at low cost.However,the massive data in IoV which are usually accessed concurrently by many users have brought great challenges to conventional IoV systems.With the above background,this paper focuses on how to efficiently store and query massive data in crowdsensing-based IoV by a distributed computer cluster.The main works and contributions are as follows:(1)Design an integrated application platform for Crowdsensing-based IoV.By researching the basic architecture of IoV and popular traffic data collection technology,as well as Hadoop cluster architecture,this paper design an integrated application platform for Crowdsensing-based IoV,which mainly contains the design of Crowdsensing traffic data gathering software on mobile terminals and the construction of cloud server cluster.The former is a piece of software developed on the Android system to achieve real-time traffic data collection by numerous users.The latter is the data center builded on Hadoop and HBase to realize the storage and querying of massive data in IoV.(2)Design a storage scheme for massive data in IoV.The scheme firstly contains a preprocessing work to exclude the missing,reduplicative and abnormal data of the original dataset.Particularly we rectify the biased geographic data by map matching and supplement clearer road information to enrich the selection criterions when querying traffic conditions.Then a kind of data storage table suitable for massive data in IoV is proposed to meet the requirements in practice.For trajectory data and data gathered by software mentioned in(1),we respectively organize column families and row keys to make data distribute orderly in HBase table so that the efficiency of querying can be enhanced.(3)Design various querying patterns for data in IoV.By taking the spatiotemporal nature of IoV data and actual need of querying into consideration,various querying patterns are designed on account of the proposed storage scheme.In this paper we can effectively realize several querying patterns to provide useful IoV information like the driving conditions of a certain vehicle,the traffic flow of a road or an area and the location of nearby vehicles of a certain mobile terminal.Finally,comprehensive experiments conducted on the platform which we designed with real data demonstrate that the querying patterns designed based on the proposed storage scheme is feasible and rational,and it also performances well in time consumption.The research is of great significance for the practical application of Crowdsensing-based IoV.