Design And Implementation Of Driving Big Data Analysis System For Internet Of Vehicles

In order to facilitate the commercial vehicle manufacturing enterprises to better improve the production process and technology and monitor the safety of vehicle operation,the mass data generated in the process of vehicle operation will be continuously transmitted back to the manufacturing enterprises for analyzing and processing after the commercial vehicles are manufactured and delivered to the users.The running data of commercial vehicles are collected by on-board sensors and sent back to the enterprise servers through communication equipment.The increasing number of vehicle terminals,high-frequency data acquisition methods and persistent storage requirements give rise to the requirement of manufacturing knowledge discovery and analysis application based on big data technology under the mass data of Internet of Vehicles.At present,there are still several key problems in the process of using big data in commercial vehicle manufacturing enterprises:（1）Static data record all production data properties of the vehicle,and the real-time driving data record the running state attributes of each part of the vehicle.It is necessary to make statistical analysis on these multi-dimensional atributes.（2）In the process of vehicle driving,GPS track changes in real time.Users need to combine with the map to visually view the dynamic fluctuation range of the vehicle.It is difficult to quickly reproduce the route and find the location through the map for the track data generated in the current driving process.（3）Multi-dimensional attribute analysis and map real-time query focus on different perspectives.Multi-dimensional attribute analysis can allow a little delay in time,but it needs more dimensions to search.Map real-time query has less attributes,but it has higher real-time requirements.Therefore,it is necessary to design a diversified storage architecture to meet the different requirements of the two parts.This paper first expounds the research background of the driving big data analysis system of the Internet of Vehicles and its significance for the commercial vehicle manufacturing enterprises,and then analyzes the requirements of the driving big data analysis system of the Internet of Vehicles from the perspective of vehicle data query and intuitive analysis of the commercial vehicle manufacturing enterprises.Then this paper describes the design of the functional architecture,technical architecture and network architecture in the system architecture,and selects the development tools that meet the needs of the system.Finally,the detailed design of each module of the system,as well as the implementation and testing of the driving big data analysis system of Internet of Vehicles are described in detail.The whole system can be divided into two parts:the data end and the Web end,in which the data side mainly realizes the diversified storage and precalculation function module of big data.The Web end mainly realizes four functional modules:vehicle state parameter analysis module,map vehicle searching module,vehicle trajectory monitoring module,and data mining interface.The vehicle state parameter analysis module is used to analyze the multi-dimensional vehicle state parameters,the map vehicle searching module is used to query the fuel consumption and driving information of the vehicle within the range described on the map,the vehicle trajectory monitoring module is to monitor the real-time distribution range,real-time state and historical running trajectory of all vehicles,and the data mining interface module is used as a secondary development interface,access to different data mining algorithms.In the data side,this paper adopts the diversified storage schemes of ClickHouse real-time storage and Kylin multi-dimensional storage,and designs the exclusive storage module for the functions with different analysis dimensions and real-time requirements,so as to meet the different functional requirements and save the storage space resources at the same time.Specifically,this paper uses Kylin precomputing to generate multi-dimensional data cube to complete the analysis of multi-dimensional attributes in different types of vehicles.In the map real-time query module,this paper uses the storage scheme of ClickHouse column real-time database,combined with Baidu Map API method,to achieve the query efficiency requirements of map real-time query module.This paper uses cluster environment to test the driving big data analysis system for Internet of Vehicles,and achieves the expected design goal.The system has been put on line in a large state-owned commercial vehicle manufacturing enterprise and applied to the actual production environment of the manufacturing enterprise.The system has realized the storage and analysis of the growing mass data of Internet of Vehicles,which meets the actual requirement of the enterprise to use and analyze the value of the big data of Internet of Vehicles.