The Representation And Reasoning Of Orientation Relation Between Moving Point Objects

As one of the most important research directions of artificial intelligence, spatialreasoning is also a popular research field of geographic information system, spatialdatabase and spatial query. The research of spatial reasoning contains many aspects,orientation relation is the closest relation to human perceptive and it plays asignificant role in describing object behaviors. However, spatial objects in real worldalways changing dynamically, and they have characters such as real-time, diversityand complexity. The research of spatial relation in dynamic environment has beenapproved generally.To some extent, there is a solid theoretical foundation in the research oforientation relation between point objects. Recently, a researcher called Rupamproposed a method to describe objects using both relative orientation relation andrelative movingorientationrelation. Nevertheless, Rupam‘s model has to model twiceto represent relation and it has ambiguity in representing.To solve the problem above, we deduce the law of mapping of orientationrelation and relative moving orientation relation based on Oriented Point RelationAlgebra (OPRA) and build the model MPOM (Moving Point Orientation-relationModel), which can represent these two relations in one model and provide thecomplete composite algorithm.Specific work is as follows:1. An overview of the latest analysis and research of orientation relation betweenpoint objects. Then we introduced the purpose and significance of this research.2. Representation of relative orientation relation and relative moving orientationrelation in MPOM.To build the MPOM, we establish coordinate system according to the moving directions of objects, abstract objects into central orientation point of minimumboundary rectangle in moving directions. Then we can tell the orientation relationbetween target object and the reference object using the tile number recording thenumber of which target object located in reference field. MPOM, takes advantage ofrelative relation two-tuple, describes moving object. And we accomplish representingtwo orientation releation in one model by mapping the two-tuple to relative movingorientation relation according to geometrical characteristic of reference object.Dividing the reference area into16tiles makes the representation more precisecompared with8tiles in Rupam‘s model.3. Give the composite operation algorism of atom orientation relationship.Based on the concept of existential composition proposed by Skiadopolous, wegive the composite definition of MPOM. According to the geometrical character ofMPOM, we propose two theorems and the proof of composite operation. On the basisof the theorems, we give the composite operation algorism and check it usingexamples.4. Design and implement the demonstration system of model representation andcomposite operation.In the model we proposed, we use the orientation relation two-tuple to representorientation relation between objects. On the basis of the mapping relation betweentwo-tuple and relative moving orientation relation, MPOM has the ability to expressboth static and dynamic orientation relations in one model which avoiding modelingtwice. Then we give the composite algorism, it could reason new relations based onexisting ones, which could study and forecast the behavior of moving objects.