The Analysis And Design Method Of Cyber-Physical System Based On Big Data

With an explosive growth of global data, the explosive growth makes the data hard to handle and save. The big data era has arrived. Big data involves in many areas, for instance, web text file, web log, Electronic commerce, social network, sensor network, astronomy, genomics, medical record and video surveillance. Therefore, Big data has to be dealt with in different fields. On the one hand, Big data era also brings new opportunity and potential value of information are hidden behind the Big data which have to be extracted by big data technology. On the other hand, Big data brings its challenge, such as how to organize and manage data set effectively.Cyber-Physical System (CPS) was first put forward by America Natural Science Fund Committee in 2006. CPS is a controllable, credible and scalable cyber physical system, which is based on perception of the surrounding environment, mixing calculation, communication and control together. Using the recyclable feedback of the calculation process and physical process, CPS can achieve integration and real-time interactivity.Meanwhile, CPS can expand and increase new function so that it can check or control physical entities in a reliable, safe, real-time and efficient way.CPS makes a comprehensive cooperation of calculation, communication and control, combines computational unit and physical unit, to achieve dynamical control and real-time perception of a complex system.This article combines big data and CPS, makes use of structural analysis and design language (AADL), establishes a model of CPS, take automatic train control system for example, analyses,expands and validate the system. This article included an analysis of big data and CPS, which contains 6v attribute in big data, MapReduce model, real-time technology and CPS based on clouds. On the other hand, this article puts forward a demand for making a model of CPS based on big data, and then use AADL to extended analysis to Spatio-temporal data model, dynamic continuous demand and error model about big data. Then, this article takes automatic train control system for example, makes use of AADL to establish the automatic train control system model and analyse it, which focus on overall system, ATC, ATP, ATS, ATO and data center. At the same time, this article establishes reasonable models of non-functional demand of the system, data flow and big data attributes. Next, this article makes a expansion of automatic train control system, mainly included: Spatio-temporal、Behavior、MapReduce and Error model. At last, this article binds all components into the system, and to validate and analyse the system. The validity of the automatic train control system can be proved by the overall model, schedulability and flow analysis. Therefore, the correctness of the analysis and design method can be concluded.