Correcting Stored RFID Data with Non-Monotonic Reasoning

Abstract

Radio Frequency Identification (RFID) technology has been held back from wide-scale commercial deployment for years due to the high level of errors in data. Despite the process of filtering the data at the edge (where RFID tags are being scanned), a significant portion of incorrect data is still inserted into the database. This incorrect data can cause inconsistency in database and can significantly influence a business logic. The errors that cause these inconsistencies include duplicate read, miss read and redundant read. While some work presented in the literature addressed the issue of cleaning RFID data, existing methods cannot solve problems if there is more that one solution or the solutions are coupled ambiguously. In this study, we present a Non-Monotonic Reasoning method that utilises two techniques to clean the RFID data stored in the database, in order to enhance its accuracy. Experimental results show, that \emph{Non-Monotonic Reasoning} can be efficiently used for cleaning RFID data, and can also obtain a higher cleaning rate when compared with the two traditional cleaning techniques alone.