Dynamic Data Fusion Algorithm Based On Context-Aware

Along with the development of network techonolgies, information services havebeen applied into the real world. There are large number of information servicesaround us, how to provide proper services in the proper time for users, so the users’satification can be increased, has become the hot research area today. In order toachive users’ acquirment of service satisfication, information service systems shouldbe adapt themselves to satisfy users. Therefore context-aware has been attractedmore and more attention.The environment of context-aware is dynamic changing, so the contextinformation collected have characteritic of uncertainty and abundance. How toreduce the ambiguity of context to ensure accurate interaction, are the research pointsof our paper. Data fusion technologies can integrate information from multiplesources, so the accurate decision can be calculated and better interaction can beguaranteed. Nowdays data fusion techonolgies have been applied into the area ofsmart home healthcare, artificial intelligence, ocean monitor etc..In the environment of smart home, our paper studys the data fusion algorithmsbased on context-aware which lack of status predication with time-varying andcontext transmission cost, leading to inaccurate data fusion results and low servicesatisfication.This paper introduces space-based context model firstly, describing the sensors’history by three aspects, context attribute, context state and situation space.Secondly, choose sensors with the maximum overall utility combining dynamicBayesian networks method. Then, related situation prediction is obtained throughdata fusion. This paper gives dynamic data fusion method in the data fusionalgorithm CFACA(Context Fusion Algorithm Based on Context-Aware), and verifiesthis method. Finally, Simulation results show the effectiveness of the CFACA.The innovative point of our paper is combine dynamic Bayesian networksmethod and context transmission cost, according to calculate the ambiguity reducing utility and context transmission cost, fuse the chosen optional attribute set with themaximum overall at each time slice and intrinsic attributes. So that more accuratestatus predication can be obtained and better users’ satisfication.