Research On Indoor Moving Objects Supporting Context Information

In recent years,the application range of geographic information science has gradually changed from macro to micro,and the spatial scale of operations has changed from the open outdoor environment to the complex indoor space.All kinds of location-related services for indoor space have received great attention.In indoor space,due to the complexity of structure and the diversity of information,more technologies for indoor moving objects are needed in production and life.At present,the study of indoor moving objects is mainly focused on dealing with the spatial and temporal information of indoor space and indoor moving objects,trying to meet the requirements of indoor applications and services.However,compared to the outdoor environment,the indoor space needs more emphasis on personalized location services,so the contextual information need to be taken into consideration of the research on indoor moving objects.In this paper,aiming at the characteristics of indoor space and context-aware requirements,we propose a data model for indoor space and some query processing methods for indoor moving objects to support contextual information.The main research work and contributions are summarized as follows:(1)After summarizing the advantages and shortcomings of existing data model for indoor space,a context-dependent double-layered data model for indoor space is proposed,including both space layer and context layer,which is focus on the integration of geometric,topological and contextual information.We introduce in the model a new method called the fine-grained partition for indoor space and give the definition of distance of indoor space with obstacles and context ontology,which add the context and preference into the model.Finally,the feasibility and validity of the modeling method are illustrated by the examples and the advantages of the model.(2)Considering the effect of contextual information on indoor path planning,a method of context-constrained indoor path query is presented.The indoor space is represented by hierarchical modeling method and directional weighted graph.The definition of indoor path length with obstacles,context-constrained property set and context constraint for path is introduced.We present a method of context-constrained indoor path query based on parallel Dijkstra algor ithm(PD-CIP)and give a method to update the path continuously.The validity and efficiency of the PD-CIP algor ithm are validated by a large number of experiments,which are compared with the existing indoor path query algorithms from several perspectives.(3)For indoor moving objects,considering the requirement of context awareness and the concept of group at the same time,a context-dependent group Top-k query method for indoor space is proposed.The definition of intra-group grouping method and a context-dependent score function for indoor space are given,and the new score function quantify the attributes of indoor moving objects,such as contextual information,by the query similarity based on vector space model.An approximate algor ithm IGCTop-k is proposed to calculate the result set of context-dependent group Top-k query for indoor space,and the algorithm is optimized by an aggregation optimization method.Experimental results show that the IGCTop-k algorithm is effective and the query accuracy has been improved.