Single versus Multi-Tone Wireless Power Transfer with Physically Large Array

TL;DR

The paper addresses the challenge of powering energy-neutral devices during the initial access phase of large-array wireless power transfer. It compares adaptive single-tone and multi-tone waveforms using a 84-antenna Techtile testbed and an energy profiler to measure RF-to-DC efficiency and END wake-up times. The main finding is that adaptive single-tone excitation yields higher harvester efficiency and shorter response times at equivalent radiated power, enabling faster pilot transmission via backscatter. This work demonstrates the practicality of single-tone strategies for initial charging in massive MIMO-like WPT deployments and points to future directions in backscatter uplink reciprocity-based beamforming and hardware-power optimizations.

Abstract

Distributed beamforming is a key enabler to provide power wirelessly to a massive amount of energy-neutral devices (ENDs). However, without prior information and fully depleted ENDs, initially powering these devices efficiently is an open question. This work investigates and assesses the feasibility of harvesting sufficient energy to transmit a backscatter pilot signal from the END, which can be then used for coherent downlink transmission. We experimentally evaluated adaptive single-tone and multi-tone signals during initial charging. The results indicate that the response time for ENDs with unknown locations can extend to several tens of seconds. Notably, the adaptive single-tone excitation shows, among others, better performance at lower transmit power levels, providing a faster response. These findings underscore the potential of adaptive single-tone signals in optimizing power delivery to END in future networks.

Figures (7)

• Figure 1: Considered three-phase operation: 1) non-coherent downlink (left), 2) 10-byte uplink pilot transmission (middle) and 3) reciprocity-based coherent downlink (right). This work focuses on the two first phases.
• Figure 2: Illustration of the real-life testbed callebaut2022techtile and energy profiler's location.
• Figure 3: Hardware architectures. The transparent blocks of the architecture were not considered during the analysis, as the primary goal was to measure the power consumption of the and the RF switch.
• Figure 4: Picture of the receive side of the measurement setup, showing two half-wavelength antennas and the energy profiler.
• Figure 5: Average power levels and corresponding efficiency levels of the harvester. The plot illustrates the relationship between average RF input- and output-power across the energy harvester. The adaptive single-tone configuration achieves, on average, higher efficiency levels.