Committee:

Dr. Manos Tentzeris, ECE, Chair , Advisor

Dr. Andrew Peterson, ECE

Dr. Gregory Durgin, ECE

Dr. Sangkil Kim, Pusan National Univ

Dr. Yang Wang, CEE

Abstract:

The objective of the presented research is to propose and demonstrate ML applications of wireless power transfer and identification technology. Several works describe the implementation of a Machine Learning (ML) strategy based on 1) the use of Neural Networks (NN) for real-time range adaptive automatic impedance matching of Wireless Power Transfer (WPT) applications, 2) the Naive Bayes algorithm for the prediction of the drone’s position, thus enhancing the WPT efficiency, and 3) the Support Vector Machine (SVM) classification strategy for read/interrogation enhancement in chipless RFID applications. The ML approach for the effective prediction of the optimal parameters of the tunable matching network, and classification range-adaptive transmitter coils (Tx) is introduced, aiming to achieve an effective automatic impedance matching over a wide range of relative distances. A novel WPT system consisting of a tunable matching circuit and 3 Tx coils which have different radius controlled by trained NN models is characterized. A proof-of-concept WPT platform which allows the accurate prediction of the drone’s position based on the flight data utilizing ML classification using the Naive Bayes algorithm is also given. A MLbased 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 high tag-ID detection accuracy. A SVM algorithm was trained using measurement data, and its accuracy was tested and characterized as a function of the included training data. In summary, this research sets a precedent, opening the door to a rich and wide area of research for the implementation of ML methods for the enhancement of WPT and chipless RFID applications.