Computer Vision-Aided RFID Tag Group Location Measurement and Reading Performance Prediction Using Graph Transformer Network With Novel Topology Graph Construction Strategy

With the increasing application numbers of radio frequency identification (RFID) multitags in recent years, the position distribution of RFID multitag has a significant impact on the reading performance of RFID system. The poor reading performance of RFID multitag will lead to the loss of important information stored in the tags. In this article, a novel synergistic topology graph construction strategy using graph attention transformer network (STGC-GATN) method is proposed to predict the reading performance of RFID multitag so as to optimize the RFID tags’ position distribution. First, considering increasingly complex electromagnetic environment, computer vision approach is involved to design and construct the RFID multitag reading performance analysis system. Then, scale-invariant SURF image matching method is adopted to effectively and accurately measure the multitags’ 3-D positions. After that, addressing the complex topology structure presented by the RFID multitag, a novel STGC-GATN method is proposed to construct the nonlinear relationship between the RFID multitag’s 3-D positions and corresponding reading distance. The proposed synergistic topology graph construction (STGC) strategy makes up for the insensitivity of graph attention transformer network (GATN) model to similar topology structures by fusing multiple distance functions, which can effectively improve the prediction accuracy of the reading distance of RFID multitag system. A large number of experiments are conducted to verify the effectiveness of the proposed methods. The experimental results show that compared with other benchmarks, the proposed STGC-GATN model can effectively forecast the reading distance of RFID multitag. The mean absolute error (MAE), mean absolute percent error (MAPE), root mean square error (RMSE), and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula> of the prediction results of STGC-GATN are 0.0307, 1.96%, 0.0383, and 0.9898, respectively. By analyzing the predicted reading distance and the topology structure, the optimal multitag 3-D position structure can be found. The found optimal structure of RFID multitag can guide the position distribution of multitag so as to improve the reading performance of RFID system.