Machine Learning Approach for Wirelessly Powered RFID-Based Backscattering Sensor System

Abstract

A RFID-based backscattering sensor system is low-cost and scalable wireless sensing technology. It takes advantage of RF energy harvesting technique, wireless power transfer and backscattering principle. It also has uncountable number of applications due to its versatility. In this paper, operation principle, architecture and machine learning technique for wirelessly powered RFID-based backscattering sensor system is presented. For the sensor tag-reading and power-delivering algorithm, machine learning techniques, such as support vector machine (SVM), artificial neural networks (ANN), and naive Bayes algorithm, are introduced with experimental verifications. The supervised SVM algorithm significantly improved reading accuracy of chipless RFID sensor tags due to superior signal classification performance of the SVM method. The ANN-based adaptive dynamic matching network for magnetic resonant wireless power transfer system improved wireless power transfer distance efficiently. Position estimation method based on the naive Bayes algorithm that is essential for smart wireless power transfer platform for wirelessly powered drones is also discussed in this paper.