Research On The Framework And Algorithm Of Uncertainty Elimination Based On Context Quality Management

In recent years,with the in-depth research in techniques,such as intelligent computing,Internet of things(IoT)and human-computer interaction,context-aware technology has got rapid development The goal of context-aware computing is to provide users with the right services in the "ever-present and ubiquitous" initiatively.There are some uncertainty problems such as incompleteness,inaccuracy and inconsistency in the dynamic and complex context-aware environment because of the wide variety and large scale of context,which lead to the systems making wrong decisions,and even lowering users' experience.Therefore,it is very important to eliminate the uncertainty of context and improve the context-aware rate.To solve the above context uncertainty problems,this thesis mainly studies QoX adaptive management based context uncertainty elimination system framework model.and multi-source continuous context inconsistency elimination algorithm based on improved basic probability assignment(BPA)of Dempster-Shafer(D-S)theory.The specific research contents are as follows:(1)For dynamism and complexity of context-aware environment,a QoX adaptive management based context uncertainty elimination system framework model is proposed,and applied the model in intelligent medical.In this thesis,the concept of QoX is first proposed as a new hierarchical comprehensive evaluation index of quality,which is based on quality of device(QoD),quality of the context(QoC),quality of service(QoS)and quality of experience(QoE).The system framework model introduces the QoX indicator and error recovery mechanism of the application layer.By the calculation,analysis and management of these quality indicators,the framework model fully improves the reliability of context-aware system.Moreover,it provides an effective method for context uncertainty elimination,which makes the context-aware system has better initiative,fault tolerance and adaptability.(2)For multi-source homogeneous and non-homogeneous inconsistent continuous context acquired from the dynamic and heterogeneous context-aware systems(CASs),this thesis put forward multi-source homogeneous and non-homogeneous continuous context inconsistency elimination algorithm based on improved BPA of D-S theory.In terms of the multi-source homogeneous continuous context inconsistency,sensor precision,membership degree and current data distance are adopted to modify the BPA.For the multi-source non-homogeneous continuous context inconsistency,the BPA is innovatively defined by three factors:sensor precision,membership degree and grey relational grade(GRG).The accuracy and fault-tolerant performance of the proposed multi-source homogeneous and non-homogeneous continuous context inconsistency elimination algorithms are further verified through experimental simulation.Based on the research of context uncertainty elimination system framework model by using QoX adaptive management,a more initiative,adaptable and fault-tolerant context-aware system framework model is built to provide support for practical operation of the context-aware system.Besides,the context-aware rate and fault-tolerant performance of the proposed multi-source homogeneous and non-homogeneous continuous context inconsistency elimination algorithm based on improved BPA of D-S theory have improved.The above researches provide an effective method for multi-source homogeneous and non-homogeneous continuous context uncertainty elimination,provide a solid calculation foundation for the context-aware system framework,and provide a powerful guarantee for the efficient operation of context-aware system.