| Summary: | 碩士 === 國立成功大學 === 測量及空間資訊學系 === 107 === The rapid growth of sensor and IoT technology brings revolutionary improvements to the quality of humans’ daily lives. As a tremendous volume of sensed data is continuously collected and distributed via mobile networks, we enjoy the luxury of using real-time data to develop innovated applications, e.g., cross-domain applications in smart cities. However, the availability of data does not necessarily mean the data can automatically fit together. Especially for applications demanding the fusion of real-time data from different sources for making prompt decisions, the comprehensive understanding about the data being used is extremely important. As current applications rarely address the “real time” issue in an appropriate manner, this research intends to develop an interoperable solution to examine the real-time properties hidden in the distributed datasets.
With the consideration of different aspects in spatial, temporal, thematic and quality information, we propose a conceptual model specifically for modelling the various aspects of the real-time properties of the GIS data. The design aims to provide necessary real-time considerations for both observations and alert information to ensure all the data acquired from different domain providers can be processed together in a standardized and interoperability way. The conceptual model includes 4 major parts: common information for observation and sensor, direct observation, derived observation and alert information. By following the conceptual model, users no longer need to create all kind of connectors to access different types of data. We demonstrate the feasibility of the proposed approach by successfully mapping the real-time open data to the conceptual model. Based on the designed elements in the standardized schema, we further propose six different working modes for real-time applications, namely, latest available information mode, valid information mode, ideal information mode, next update information mode, historical information mode and alert information mode. As every mode is designed according to the considerations of real-time, data type and quality, it can be directly used for smartly and automatically presenting only the datasets that are reliable and helpful for users’ decision making.
Following the defined modes, we further propose a status-aware interface for visually presenting the various aspects of differences among the selected datasets. The interface design is augmented with cartographic knowledge to enable users to visually distinguish data with different real-time properties and avoid wrong decision making. In addition to the map interface, analyzed outcomes are also presented with textual or widgets to provide additional guidance. Map symbols are designed for showing data without calibration, time differences to the current time, observation values, different data source and observation value. From the perspective of showing the additional information for real-time data, different visual aids are included: mode selection and filter, time sequence diagram, warning message box, time statistic gauge, indicator and map legend.
As real-time GIS applications often operate under extremely high time pressure, the proposed standardized approach proves to be able to provide a feasible solution to improve the interoperability of sensor observations in real-time GIS applications. Only after ensuring the necessary information for all the observations is available, can we successfully take the best advantages of the abundant volume and a variety of domain observations in the internet-based exchange environment.
|
|---|
