Research On Self-Adaptive Context Pre-processing Model In Context-aware Middleware

With the development of internet, sensors and mobile devices, pervasive computing is more and more popular. As a branch of pervasive computing, context-aware computing becomes a hot research field, including middleware, mobile computing, human-computer interaction, embedded technology, sensor technology, network technology, etc. The development of context-aware middleware can solve the problem of heterogeneous and cross-platform, provide a integrated context aware application platform.Context-aware middleware exacts the common properties into middleware from context-aware computing, shielding the heterogeneity of context sources and improving the reuse of context information, releasing the burden of developers, providing more accurate advanced context information and more intelligence reasoning technology, and ultimately improving the user experience.In the real world, there are fulfilled of contexts, which may be poor quality so as to reduce the efficiency of application. Thus, it is of vital importance to provide effective and efficient reasoning in middleware. So we need to preprocess the raw contexts. From the statistics, we know that the data preprocess accounts for 60% of data discovery. The data preprocess can improve data quality, make the middleware more efficiency.Due to the conclusion above, this paper proposes a Reflection-based Multi-Agents Context Pre-processing Model (RMACPM) in term of General adaptive Context Aware Middleware Model (GACAM). This is a autonomous, dynamic and intelligent context pre-processing model, which is composed of filtering module, converting module, fusion module and monitor module. Through coordination and interaction of these four modules, RMACPM achieves a self-adaptive pre-processing model.Finally, combining with the projects in lab, we designed and implemented this pre-processing model. Through the display of classical scenarios, we get the conclusions.