Real-Time GIS Data Model Supporting Dynamic Data Management And Spatiotemporal Porcess Simulation

Geographic phenomena in the physical world is constantly changing. In order to develop a beneficial coping strategy when some man-made emergencies or major natural disasters occur, the people always want to know the latest states of these geographic phenomena, and even their trends. The traditional way to obtain spatial data can not meet the needs of such time-sensitive. With the wide and in-depth application of LBS (Location-based Sevices) and various sensors, spatiotemporal data acquisition becomes more efficient and convenient, thus, a large number of high time efficiency spatiotemporal sequence data is producted. However, traditional static GIS spatial data models only express the spatial distribution of geographic phenomena, and they do not consider the temporal characteristics. Temporal GIS spatiotemporal data model increases the time dimension description, and temporal change process of geographic phenomena can be expressed. The main work of existing spatiotemporal data models is to manage massive historical spatiotemporal sequence data and relationships between the data, however, they lack of the performance in real-time and simulation.This dissertation analyzes the structures and rules of spatiotemporal change process of geographic phenomena in the physical world based on the existing spatiotemporal data models, then a real-time GIS data model is porposed using object-oriented technology and combining with the characteristics of sensor observations and Agent simulation. So, the real-time GIS data model has usability and versatility, and it can support dynamic data management and spatiotemporal process simulation. The model lay a solid foundation for the development of real-time GIS, and it also provide technical support for the construction of smart city.After referencing the conceptions and characteristics of real-time system and real-time database, and analyzing the development process form traditional static GIS to temporal GIS, this dissertation gives a conception of real-time GIS and divides the real-time GIS into hard real-time GIS and soft real-time GIS, and the soft one is the object of study in this dissertation. Real-time performance is the most essential characteristics of the real-time GIS. The spatiotemoral semantic structure of this dissertation is that time semantic located on space semantic and thematic semantic. This dissertation studies the existing spatiotemoral data models in the form of different periods and classifications, and analyzes the advantages and disadvantages of existing spatiotemporal data model, it provide a reference for studying real-time GIS data model.In real-time GIS, geographic object is the main subject of spatiotemporal change, its state data comes form real-time observation using sensor. In view of this, a real-time GIS geographic object storage model is proposed centering on sensor object and geographic object. In the model, sensor object is composed of parameter state and observation state, and geographic object is constituded by a plurality of sub-state, the data of each sub-state are from observation of sensor. This dissertation expresses the relationships between various elements of the model, class structure, and database storage structure.The nature and rule of spatiotemporal change process are analyzed, and the meaning and relationship between various elements are clearly expressed. On the base of real-time GIS geography object model, increasing spatiotemporal process, event type, event, observation, geography object simulated state and other elements, proposed a supporting spatiotemporal process expressing real-time GIS data model. In which, spatiotemporal process is used to describe changes of spatiotemporal scene, including geography objects and events. Geographic object can generate event when some conditions are satisfied, and the event can also drive geographic objects when some conditions are satisfied, thus, spatiotemporal process can be better expressed and simulate. The event type includes the conditions of that geographic object generates event and/or the conditions of that event drives geographic object, and determine the relationship between geographic objects and events. This dissertation describes the entire process from concept definition, relationship description, logical expression, to data storage, it lays a foundation for implementing the real-time GIS data model.The ability of the real-time GIS data model in simulating spatiotemporal process is limited. This dissertation studies the relationship that the elements of the real-time GIS data model map to Agent, it can improve the ability of the real-time GIS data model in simulating spatiotemporal process using Agent on system dynamics performance. The geographic object is mapped to the Agent, and it is considered as an Agent. The conditions of event type is mapped to simple reaction type Agent condition-action rule, or it is mapped to careful reflection Agent repository, for a better combination between real-time GIS data model and ABM.The dissertation use three experiments to verify the ability of the real-time GIS data model in managing dynamic real-time data and simulating spatiotemporal process. Firstly, a data model is designed using real-time monitoring location and picture data of Wuhan taxi, and the experiment storaging and expressing the move objects is implemented. Secondly, a data model is designed using PM2.5monitoring data of9national monitoring stations in Wuhan, and the experiment storaging and expressing the situ objects is implemented. Finally, a spatiotemporal process scene, waterlogging in Wuhan Hangkong overpass intersection (first-level waterlogging risk point) impact surrounding traffic, is designed, and the experiment expressing and simulating the spatiotemporal process is implemented using Agent.