The Construction And Combination Technologies Of Cloud Services For Big Data-driven Intelligent Transportation Systems

In the process of intelligent transportation systems development, the big data collected by GPS probe vehicles, microwave detection traffic sensors SCATS coil, namely multi-source mass sensor data, has become an important resource for research in intelligent transportation system. So people proposed big data-driven intelligent transportation system, which is mainly through the massive multi-source heterogeneous traffic data analysis and processing, to provide data-based solutions for intelligent transportation systems. However, in the current big data-driven intelligent transport systems, there are still many difficulties, the first is the traditional system is not only the deployment of a long cycle, high maintenance costs, and, with the increase in the amount of data, increasing the strength of the business and the calculation complexity increases, it is difficult to meet the growing demand by increasing the horizontal expansion. Second, higher coupling between system modules, it is difficult to achieve reuse and sharing of modules. In addition, the system data for each data source requires customized treatment, specificity, not only affects the efficiency of program development, but also resulted in waste of resources. Based on the problems in current large data-driven intelligent transportation system, in this paper, construction and combination technology of big data-driven intelligent transport cloud has been studied. The first is to build traffic cloud platform based on Open Stack to provide secure and reliable infrastructure services through Open Stack for the system, which can be deployed via an inexpensive machine, reducing the construction cost of the system. In addition, traffic cloud platform flexible elastic deployment capabilities, not only reduced the deployment cycle of the system, but also can meet the growing mass transportation data processing needs in future by increasing the number of Open Stack compute nodes at the same time.Then, based on the traffic cloud platform to build a RESTful Web service-based traffic cloud service,through each module in the system is converted to transportation cloud services, reduce the coupling between modules and improve reusability and sharing of each module; And finally, designed a service composition model, by combining simple services to complete complex transportation business processes.Based on the above traffic cloud service combination technology,This paper realized the two kinds of application systems. First, construct the map-matching algorithm, virtual guidance screen road network planning algorithm and coordinate conversion algorithm to the corresponding traffic cloud services, Through traffic cloud service combination technology to design a kind of location service based dynamic virtual induction screen system, it can be seen through the experiment, compared with the traditional induction method, this method can effectively improve the efficiency of induction.Then through transportation cloud services based on D-S evidence theory adaptive fusion algorithm and load balancing algorithm of Quadtree designed parallel processing scheme based on traffic cloud services, The experimental analyzed the relationship between the data processing time and the virtual machine computing nodes, the experimental results show that the parallel scheme based on traffic cloud service composition can effectively improve real-time massive multi-source heterogeneous data integration.