Feature selection for RFID tag identification

We present a multi-objective optimization (MOOP) based feature selection technique for radio frequency identification (RFID) where a tag is identified by matching a set of unique characteristics measured from the tag to previous stored copies in a database. The aim of this paper is to select the most effective characteristics for tag identification. Different application scenarios require different levels of security and demand different false match rate (FMR) and false non-match rate (FNMR). To handle those two conflicting objectives in feature selection, we formulated it as a MOOP problem that generates a set of best possible FMR and FNMR performances a system can achieve with different feature combinations. Experiment results show that the proposed technique can provide a broad view of the effectiveness of the system permitting a system designer to meet specific security requirements for a given application scenario.