Intelligent RFID Tag Detection Using Support Vector Machine

Abstract

RFID Tag detection/recognition is one of the most critical issues for successful deployment of RFID systems in diverse applications. The main factors influencing tag detection by RFID reader antenna include tag position, relative position of reader, read field length, etc. In this paper, we analyze the characteristics of tag detection for a carton box object on a wooden pallet by an experimental approach based on tag signal strength, and we propose a method for predicting detection related directly to the strength of tag signal using an intelligent machine learning technique called support vector machine (SVM). The use of the proposed method is able to save time and cost by quick prediction of tag detection. Extensive experiments showed that the proposed approach can predict tag recognition for a carton box object with an accuracy at 95% for various reader heights and read field lengths. The proposed approach is effective for determining the best tag detection influencing factor conditioned on the target object with the help of detectability prediction.