Design And Implementation Of Context-aware Service System Based On Vehicle Network

With the rapid development of semantic web, the semantic web has been applied to many fields such as vehicle network. The problem is that RDF data is growing rapidly year after year. The tools of traditional ontology query and reasoning have been difficult to compute the ontology including a large number of instances. The limit of memory do not only affect the efficiency of the query, but also can not query the ontology including a large number of instances. Secondly, the structure of the table need to be designed while using relational database. There are some differences between the structure of the table and RDF data, so the conversion between two structures can affect computational performance.Therefore, it is necessary to study how to implement a large-scale storage model and reasoning algorithm based on distributed database.Based on previous researches, this paper combines the existing storage strategy and the characteristics of HBase, researches how to use row-key, column name, and data units in HBase to store the ontology.This paper proposes a storage model of semantic data based on HBase,and designs a inference algorithm based on ontology and user-defined rules. A context-aware system based on vehicle network is designed and implemented. The main work of this paper is as follows:(1) A storage model based on HBase is designed. The ontology data is stored separately from the instance data, and the instance data is divided into classes. A reasonable storage structure is designed to ensure retrieval efficiency by using the default dictionary characteristics of the row-key in HBase. The inference algorithm is implemented by using the HBase Java API. The algorithm of querying eight kinds of triple pattern is implemented, and the algorithm of inferring user-defined rules and three kinds of relation including subClass, subProperty,inverseProperty.(2) A context-aware system based on vehicle network is designed and implemented. The architecture of the system is introduced. After the ontology, rules and inference are detailly analyzed and designed, a specific implementation plan is given.(3) The test of function and performance on the system is performed.The result shows that the system can complete the inference of large-scale data, and verifies the effectiveness of the storage strategy and inference algorithm.