Abstract
This work describes the implementation of a machine learning (ML) classification strategy for read/interrogation enhancement in chipless Radio Frequency Identification (RFID) applications. A novel ML-based approach for classification and of detection tag IDs has been presented which can perform effective transponder readings for a wide variety of ranges and contexts, while providing tag-ID detection accuracy of up to 98.3%. Four tags encoding the four 2-bit IDs were inkjetprinted onto a flexible low-cost PET substrate and interrogated without cross-talk or clutter interference deembedding at ranges up to 40cm, with different orientations and with and without the presence of scattering objects in the vicinity of the tags and reader. A Support Vector Machine (SVM) algorithm was then trained using 259 measurements and its accuracy was tested and characterized as a function of the included training data. Finally, the excellent performance of the approach, displaying reading accuracies ranging from 89.6% to 99.3%, is reported. This effort sets a precedent, opening the door to a rich and wide area of research for the implementation of ML methods for the enhancement of chipless RFID identification applications.

This publication has 11 references indexed in Scilit: