Distributed Database Design for Mobile Geographical Applications

Manhoi Choy, Mei-Po Kwan, Hong V. Leong
2000 Journal of Database Management  
Advanced Traveler Information Systems (ATIS) require e cient information retrieval and updating in a dynamic environment at di erent geographical scales. ATIS applications are useful in yielding a better utilization of the limited costly transportation arteries and providing value-added traveler information. Many A TIS applications are built on the functionalities provided by Geographical Information Systems (GIS), which often cannot meet extra requirements like real-time response. We i n
more » ... ate GIS-based systems in ATIS and propose a system architecture based on GIS and distributed database technology. Issues on data modeling, data representation, storage and retrieval, data aggregation, and parallel processing of queries are discussed. This paper introduces a distributed system architecture for ATIS based on recent t e c hnology presents new data models for information representation proposes data shipping for e cient query processing and function shipping in reducing communication overhead exploits network of computers for solving complex problems more timely and incorporates privacy protection for sensitive data. He is supported by g r a n ts HKUST6071/97E and DAG96/97.EG25. Mei-Po K w an's current address is: Department of Geography, Ohio State University, Columbus, OH 43210, USA. Hong Va Leong's current address is: Department of Computing, The Hong Kong Polytechnic University, Hong Kong. Authorship is in alphabetical order. traveler decision making can be integrated through geo-referencing. GIS are also highly exible in manipulating spatial objects and distance according to rules, and di erent scenarios can be simulated to test the \what-if" cases. GIS operations can help to de ne individuals' accessibility and mobility in space and time (Kwan, 1998b Kwan & Hong, 1998 . It is particularly helpful in modeling human travel behavior in the context of ATIS (Yim, 1996) . Kwan (1998a) has suggested that the requirements of a GIS for modeling human travel behavior in an ATIS context. The system needs to handle not only route information, but also a large amount of explicit spatial information about activity locations in the environment. Both activity locations in the objective and subjective e n vironment (preferred activity locations, preferred routes and familiar areas) would be important i n modeling travel behavior. A GIS-interfaced computational process model (CPM) called GISICAS is developed and implemented for activity s c heduling behavior in ATIS (Kwan, 1998a). GISICAS contains a comprehensive geographic database in ARC/INFO format to supply detailed spatial information about the potential activity opportunities. It models how the traveler processes the spatial information supplied and makes complex decisions in real-time using spatial search algorithms. The spatial information is stored in a comprehensive geographic database containing a detailed street topological network and activity locations. The commonly used GIS data models, however, are not without problems. For example, the raster data model divides space into regularly shaped and sized pixels, whereas the topological data model subdivides space into irregularly shaped regions, links and nodes (Frank, 1992) . None of them, however, represents tra c movement a n d interaction very well and the problem of connectivity is not taken into account. Although some GIS packages such as TransCAD (Caliper Corportaion, 1996) and ARC/INFO (ESRI, 1996) implement transportation functions like routing algorithms in the topological data model, they are not without problems. First, the link-node structure is basically planar and would not distinguish an intersection with an overpass which does not cross at grade. This would induce problems for routing unless additional structures in the data model are added. Second, the topological model does not replicate how h uman perceives the street network. We usually do not think of the street network as segments of links with intersection, but more as the street as a whole. As such, the topological data model is not a natural navigable database . In addition, ITS applications require operations at both regional level and local level. Information may need to be transmitted between di erent levels of modeling. Mainguenaud (1995) used an object-oriented (OO) approach with graph theory to show h o w hierarchical relationships can be incorporated into networks. His approach a l l o ws single nodes (Master nodes) or lines (Master edges) to represent abstractions of sub-networks of the original network. These Master nodes/edges would thus represent subnodes and edges at a di erent scale. Although this approach is very useful, dynamic information needed in ITS is not dealt with. OO data modeling as an alternative for spatial databases was discussed by Worboys et al.
doi:10.4018/jdm.2000010101 fatcat:yu2oefoiwvblvj4sy4fgn4rn2u