Research And Implementation Of The Construction And Query Of The Knowledge Graph Of The Internet Of Vehicles Based On PostgreSQL

With the increasing popularity of Internet of Vehicles technology,real-time operating data of commercial trucks have become easy to obtain.In order to make full use of the data of Internet of Vehicles,it is necessary to comprehensively analyze the constraints and associations between various vehicle conditions and accurately mine the value of data of Internet of Vehicles.However,the relationship between the data of the Internet of Vehicles is complicated.In addition to the structured data collected by the on-board sensors in real time,it also includes a lot of mechanism knowledge information,and the data have strong time-series characteristics.How to model the data of Internet of Vehicles is a big challenge.The knowledge graph of Internet of Vehicles can provide strong support for the treatment of these problems.The general construction method of the domain knowledge graph is relatively mature,but when it is combined with a certain domain,it is necessary to fully consider the characteristics of the data.The data of Internet of Vehicles have characteristics of obvious time series,multi-modality,and strong real-time performance,and contain formulas and rule mechanism knowledge.In general,the research on the knowledge graph of the Internet of Vehicles faces the following problems:（1）The lack of a unified modeling representation method for the data of Internet of Vehicles and mechanism knowledge,and it is difficult to effectively express the data and mechanism knowledge in the knowledge graph.（2）It is difficult to reflect the characteristics of the time series of the data of the Internet of Vehicles,and the static knowledge graph structure cannot reflect the time sequence correlation.（3）There are performance bottlenecks in the application of the knowledge graph,and performance problems will occur when the knowledge graph structure and data are frequently inquired.In response to these problems,this paper uses the knowledge graph technology to model the data and knowledge of the Internet of Vehicles in a unified way,and uses the LSTM artificial neural network to represent the time series data,thereby completing the updating of the correlation degree in the knowledge graph.Finally,we use multilevel performance optimization methods like adaptive optimization technology,flame graph drawing and analysis,database GUC parameter optimization,explicit change of query statement execution plan to provide more efficient supports for the updating and retrieval of knowledge graph from the database level.The main works of this paper are summarized in the following three points:1.This paper gives the construction scheme of the data and mechanism knowledge of the knowledge graph of the Internet of Vehicles.This paper divides entities according to the characteristics of the data,and divides relationships into different types at a finegrained level,and sets relevant attributes for entities and relations.This paper uses mechanism entities to associate certain entities with semantic information such as mechanism knowledge,and completes the unified modeling representation of the knowledge graph.2.This paper proposes a knowledge graph updating method that considers the time series characteristics of real-time data.The real-time data are divided into vehicle body types according to the engine and wheelbase type of vehicles,and the LSTM neural network is trained for the time series data of each type of vehicle body,the time series relationship is expressed,and the correlation coefficient is calculated on the obtained output set.The calculation and updating of the relevant informations of the knowledge graph provide a more reliable data association basis.3.This paper provides a PostgreSQL-based storage solution and performance optimization solution for the knowledge graph of Internet of Vehicles.The data and structure of knowledge graph are stored in PostgreSQL,and a series of database performance optimization methods are used to improve the retrieval and updating efficiency of the knowledge graph,and provide more powerful performance support for upper-level applications from the database level.