Research Of Context Inconsistency Detection And Resolution

With the rapid development of the information technology, the pervasive computing becomes the natural trend of the computing technology development. Because of the rapid changing of the pervasive computing environment, the computing systems in it must be able to aware these contexts automatically and modify themselves. So, as the developing of the pervasive computing research, context-aware technology has become a hotspot. The context- aware applications based on the context information can't guarantee that the context data from the physical sensor is fully accurate and the rules of inference are absolutely right, and then the context inconsistency problem is raised. How to ensure the reliability and accuracy of context information, in other words, the research of context inconsistency problem, has become a very important part of context-aware research.This research was carried out against the context inconsistency problem, and mainly studies the solutions that can detect and resolve it effectively. Specific including:(1) Describe the context modeling solution based on semantic matching in detail. Include the context's representation, the definition of rules, context matching, and other important concepts, as well as a context inconsistency detection and elimination process based on a series of algorithms. Ultimately form a set of context inconsistency detection and elimination methods.(2) Give 4 basic algorithms to resolve the context inconsistency, including the algorithm which discards all the inconsistency contexts, the algorithm based on the number of inconsis- tency contexts which are eliminated, the algorithm based on the occurrence rate of an incon- sistency context, and the algorithm based on the correlation of the contexts. Through the experiments'results, this paper explains which algorithm is better and their usage. (3) Based on the 4 basic resolution algorithms, this paper gives the expansive sorting algorithm. Experiments'results show that the expansive resolution algorithm effectively increases the accuracy of context inconsistency elimination, which can provide context-aware applications more reliable and accurate context information.(4) Present a feedback mechanism based on confidence database for context incon- sistency detection and resolution. Through the statistics and analysis of the historical data on the confidence database, control the context inconsistency detector and eliminator, and make them more efficient to deal with the context inconsistency problem. And this paper verifies this prediction through experiments.