The Analysis And Design Method Of Big Data-Driven Vehicular Network Cyber-Physical System

Vehicular Network Cyber-Physical System played an important role in the development of intelligent transportation. It changed and influenced the traffic conditions by computing, communication and control technology. With the data growing, as well as the continuous development of big data technology, applications of big data in the Vehicular Network Cyber-Physical System become more and more important. The potential value of the big data had a great help to improve Vehicular Network, and improve the user’s experience. However, the explosive growth of the data volume makes the data processing and storage were facing challenges.To meet the characteristics and requirements of big data driven Vehicular Network Cyber-Physical System, such as volume, variety, velocity, veracity of big data and complexity、real-time and spatio-temporal requirements of Vehicular Network Cyber-Physical System, an architecture of Vehicular Network Cyber-Physical based on cloud computing and real-time processing model and data model and data storage model of big data were proposed. According to the data processing of big data of vehicle network, a batch processing model which based on MapReduce was analyzed. Considering the real-time requirement, a real-time big data processing model based on Storm and a large data parallel processing model based on Hadoop were proposed.A method based on SoS(System of System) and AADL(Architecture Analysis & Design Language) was proposed for analysis and design of big data driven Vehicular Network Cyber-Physical System through the research into SoS theory and AADL, combined with the properties of big data and Vehicular Network Cyber-Physical System. The hardware and software models of the Vehicular Network Cyber-Physical System were made by using AADL. Physical model and physical behavior of the Vehicular Network CPS were made and the physical model and behavior were simulated by using Modelica. Non-functional properties of the system and characteristics of big data were described by extended AADL, and corresponding code were generated. Through end to end flow latency analysis and testing and scheduling simulation, proved that the method can detect the potential problems of the system in the early stage of system design and make correction in time, and efficiency and availability of the proposed method for analysis and design of big data driven Vehicular Network Cyber-Physical System were verified.