An active multidatabase system with function of spatial and temporal context recognition for mobile computing environments

在本研究中，我们实现了一个具有时空上下文识别功能的主动移动多数据库系统。该系统根据用户的时空背景，通过动态自动集成异构数据库，为用户创建新的应用领域，提供新的信息，提高各个数据库的重要性。本研究设计了一个主动移动多数据库系统的元级系统架构。这种金属级系统架构实现了一种新的基于上下文的金属级数据集成机制。上下文相关的数据集成意味着通过我们的计算模型集成遗留数据库，以测量相等性、相似性、空间和时间关系。我们的计算模型通过组合与用户上下文相对应的计算模型来动态计算遗留数据库之间的各种关系。我们将该模型实现为金属级的混合计算模型。我们还设计了一种机制，通过将我们的模型应用于移动计算和主动数据库技术来识别用户的时空背景。这种机制使得根据用户的时空上下文动态、自动地集成异构数据库成为可能。此外，我们已经实现了一个实验系统，并进行了几个实验来阐明我们的系统的可行性和有效性。(1)从数据集成检索结果的准确性、实现开销和性能三个方面验证我们的用户时空背景识别机制的可行性和有效性。(2)将主动数据库技术集成到金属级系统中，评估系统实现开销和性能的实验。

In this study, we have realized an active mobile multidatabase system with spatio-temporal context recognition functions. This system is used to create new application fields, provide new information with the users and enhance the importance of each individual database by dynamically and automatically integrating heterogeneous databases according to a user's spatio-temporal contextIn this study, we have designed a metalevel system architecture for an active mobile multidatabase system. This metalevel system architecture realizes a new context-dependent data integration mechanism in the metalevel. The context-dependent data integration means that legacy databases are integrated by our computational model for measuring equality, similarity, spatial and temporal relationships. Our computational model dynamically computes various relationships among legacy databases by combining computational models corresponding to user's contexts. We have realized this model as the hybrid computational model in the metalevel.We have also designed a mechanism for recognizing a user's spatio-temporal context by applying our model to mobile computing and active database technologies. This mechanism makes it possible to integrate heterogeneous databases according to a user's spatio-temporal context dynamically and automatically.Furthermore, we have implemented an experimental system, and performed several experiments to clarify feasibility and effectiveness of our system.(1) The experiments for clarifying feasibility and effectiveness of our mechanism for recognizing a user's spatio-temporal context in terms of accuracy of retrieval results of data integration, the implementation overhead and performance.(2) The experiments for evaluating implementation overhead and performance of our system in integrating active database techniques into the metalevel system.