Deep RNN-Based Channel Tracking for Wireless Energy Transfer System

In this article, we study channel tracking for a wireless energy transfer (WET) system. This problem is practically very important, but challenging. Regarding time-varying channels as a sequence to be predicted, we exploit the deep learning technique for channel tracking. Particularly, by constructing a recurrent neural network (RNN) architecture based on long short-term memory (LSTM) and feedforward neural network (FNN), we develop a novel channel tracking scheme for the WET system. This scheme sequentially estimates the channel state information (CSI) at the ET based on the previous CSI estimates and the harvested energy feedback information from the ER. Numerical results demonstrate the superior performance and effectiveness of the proposed scheme.

I. Introduction

Radio frequency (RF) wireless energy transfer (WET) is a very promising technology for energy-constrained networks and it has broad applicability to wireless sensor networks, Internet-of-Things (IoT), etc., [1]–[6]. To overcome the short distance and low efficiency of the WET due to the propagation loss, in the WET systems, it is common to transfer the RF energy using the multiple-antenna technique, e.g., the energy beamforming [1], [2]. To transfer the RF energy as much as possible via the energy beamforming, the channel state information (CSI) must be perfectly or accurately known at the energy transmitter (ET), which, however, is practically very challenging due to limited processing capability of the practical energy receiver (ER) (or energy harvesting circuit) [3], [4].