Using Design of Experiment to Evaluate the Performance of a Radio Frequency Identification (RFID) Inventory Reader at the item level

Libraries are fast growing adopters of Radio Frequency Identification (RFID), as the technology promises to relieve repetitive strain injury, speed patron self-checkout, and make possible comprehensive inventory. Hevner et al (2004) resonates that researchers should demonstrate the utility, quality, and efficacy of a design artifact using rigorous evaluation methods. Despite the sound arguments on the importance of evaluation, there is a surprisingly scarce number of conceptual and empirical research on RFID performance testing in libraries. This absence of a methodology to evaluate performance of the RFID library system could suggest that the research community has neglected testing in this environment. Testing in other environments such as the supply chain has indicated that accuracy rates are contradictory to competing and misleading claims of 100% read rate alluded to by RFID vendors. As argued by Zoch and Ferguson (2004) test cases/processes should follow logical steps to identify those variables that would most affect a particular client's simulated RFID model. As a result, the purpose of this paper is to propose and execute a methodology to investigate performance of RFID library inventory reader, with the intent to investigate factors (or an interplay of factors) that may affect the utility of the reader. The methodology encompassed a design of experiment approach that investigates four independent factors: read angle (A), read distance (B), tag location (C) and shelf material (D), and one dependent factor, read rate. The findings suggested that read angle, read distance and tag location had a significant impact on the performance on the inventory reader. Initial findings on shelf type (wood or metal) suggest that this factor had an impact on read rate of the inventory reader. It is anticipated that the findings may help researchers and library practitioners to understand and further investigate methods to optimize the performance of the reader.