Research On Processing Approach To Missing Detection And Error Detection Of RFID Complex Events

At present, with the development of RFID technology, it makes this technologypopular in many application areas, but because of the restriction of the RF signal, theRFID raw data may have errors when readers collect data, it causes the unreliabilityof data, which mainly include negative reading and positive reading errors,these irresponsible data will bring mistakes to the upper layer application, it restrictsthe development of RFID technology, so the unreliable RFID raw data cleaning isnecessary. But most of data processing for unreliability uses the original data cleaningmethod to remove the error data. But these methods are not able to completelyeliminate the unreliability in the initial data stream. This unreliability willmislead RFID middleware generation not exist complex event or omit the complexevent which should happen-so produce complex event error detection and complexevent missing detection. This paper focuses on how to correct the errors of high-levelcomplex events while caused by underlying data unreliability.We studied a large number of RFID raw data unreliability in practical applications,and analyzed how can the unreliability affect the high-level complex event synthesis.Base on these analysis, we raised a method to handle the missing detection and errordetection problem.In this paper, we raised a method to handle the complex event missingdetection—TCSC: the method compresses the raw data which come from readers, toreduce the data volume in middleware; Raised a kind of logic relation model ofregional events, abstract the positions of readers according to the application and pathbetween different positions and time constraints, finally form a reader distributiongraph; Find out the complex event probability table based on statistics of historicaldata. And we process the compressed data according to this graph, to find out allpossible complex events within the time limit, finally, according to the complex eventprobability table, find out the event which has max probability within the time limit.Although complex event error detection is not so common as complex eventmissing detection, but the influence of this problem on enterprises’ and institutions’important events detection will be serious. According to this situation, this paperraises a processing strategy for the important events error detection—HDP: it is divided into three levels according to the function.1. The important object’s historicaldata cache layer, after the important inquires registration, this layer stores the historydata of the objects of the important inquires dynamically; The time domain divisionlayer, it divides the data which are stored in the first layer according to time, and getlocation information of the object.3.The matching degree calibration layer, use thelocation information of the object come form the second layer to calculate thematching degree on space between the different important object, finally compare thematching degree with the system threshold, then eliminate the error detection.