The Design And Development Of Spatio-temporal Relation Analysis And Query Component

The research of the Spatio-temporal Reasoning was began in the 1970's.It applies techniques and methods of the Artificial Intelligence in modeling,describing and expressing spatio-temporal objects. It has been a hot subject inthe research of the Artificial Intelligence.The work of this paper is a part of the national 863 high-tech project"Web Intelligence Applications Development and Runtime Platform", whichwas developed by the lab of knowledge engineering of Jilin University. Inmost applications, the support for the space (time) data is indispensable, sohow to handle and manage the spatio-temporal data is also one of researchtopics of this project. We designed and developed a spatio-temporal relationanalysis and query component based on the Spatio-temporal Reasoning.First, we have discussed the formalized describing method ofspatio-temporal relations.Spatio-temporal relations are usually described in mathematics or logicmethods. Giving the formalized semantic description of spatio-temporalrelations, its purpose is to provide a formalized tool for the spatio-temporalanalysis and query. We have analyzed the following nine qualitative modelsof spatio-temporal relations: the Interval Algebra model, the RCC model, theN-Intersection model, the Broad Boundary model, the QTCN model, theTaper Direction model, the Well Direction model, the Distance model and theSize model. We have also studied the describing method for the informationcombining time with topology, distance and size.We can find from these nine models that, many models divide the spaceinto several parts according to the object, and may deduce relations betweenobjects using intersections of their parts. So that we presented a formalizeddescription method of spatio-temporal relations based on definite partition ofthe space: the K-Frame model. In the K-Frame model, all parts of the spacedivided according to the object is called the Partition of this object;anintersection of two partitions form different objects is called a Binary-Pair;allBinary-Pairs needed for determining a kind of spatio-temporal relation iscalled the Frame of this relation;every value of the Frame is called a Case,thereinto the value that can be mapped to basic relations is called Valid-Case.In this model, to define a spatio-temporal relation we only need to give thebasic relation set, the Partition method, the Frame set, and the mapping tablebetween Valid-Cases and basic relations.Next, we have discussed the design of the spatio-temporal relationanalysis and query component.The spatio-temporal relation analysis and query component wasdeveloped based on the MVC structure, including the Spatio-temporal Model,the Deducing Control, and the User Interface. The description, reasoning anddisplay of spatio-temporal relations are in different modules, so that thededucing process is no longer related with concrete models. We defined aninterface criterion for spatio-temporal relations, so any model that matchesthe interface definition can be used in this component.There are two main functions of the spatio-temporal relation analysis andquery component: determining spatio-temporal relations between two objects,and searching target objects under given conditions. In addition, we haveprovided a model-choosing tool based on the Evidence Theory to helpcustomers choose the appropriate model.Again, we have discussed the spatio-temporal algorithm used in thiscomponent. This algorithm is based on the K-Frame model. The same part ofthe determining process and the related data structure is regard as the mainbody of the algorithm, and encapsulated in the parent class;the different partis regard as the variable body of the algorithm, and realized in the sub class.The algorithm combines two parts through dynamic binding, and keeps theformal consistency. We have given the detail of the parent class and eachconcrete model class.Finally, we have discussed applications of the spatio-temporal relationanalysis and query component in the agriculture. The spatial information isvery important in the agriculture, but one shortage of current variousagriculture application systems is the weak capability of handling spatial data,so the component just makes up this shortage.The work of this paper has importance meaning for applying theSpatio-temporal Reasoning into applications of the Geography InformationSystem. And there is still a lot of further research to complete in the future.