Time-Division Energy Beamforming for Multiuser Wireless Power Transfer With Non-Linear Energy Harvesting

Ganggang Ma, Jie Xu, Ya-Feng Liu, Mohammad R. Vedady Moghadam

TL;DR

This paper introduces a novel time-division energy beamforming strategy for multiuser wireless power transfer (WPT). By moving beyond the simplistic "linear" model of energy harvesting and accounting for the real-world non-linear "convex-concave" nature of RF-to-DC conversion, the authors demonstrate that time-sharing different beamforming patterns can significantly increase the minimum energy harvested across multiple users.

What Problem Does This Paper Address?

In Wireless Power Transfer (WPT), an energy transmitter (ET) uses multiple antennas to "beam" radio frequency (RF) signals toward energy receivers (ERs). Traditionally, researchers assumed a linear energy harvesting (EH) model, where the harvested direct current (DC) is simply proportional to the received RF power. Under this assumption, the best strategy is "multi-beam" transmission—sending a steady stream of energy to all users simultaneously.

However, real-world hardware (specifically the diodes in rectifiers) is highly non-linear. The relationship between input RF and output DC follows a convex-concave curve:

  1. Convex Region (Low Power): Conversion efficiency is very low. Increasing the input power slightly leads to a disproportionately large jump in DC output.
  2. Concave Region (High Power): The hardware begins to saturate. Increasing input power yields diminishing returns.

If users are in the convex (low-power) regime, the traditional multi-beam approach "spreads the butter too thin," resulting in poor conversion efficiency. This paper addresses how to optimize beamforming when these non-linearities are taken into account.

Methods

The authors propose a Time-Division Energy Beamforming framework. Instead of a single static beamforming strategy, the transmission time is divided into slots. In each slot, the ET can use a different beamforming matrix (which might target different groups of users with higher intensity).

Key components of their approach include: * Sigmoid-based EH Model: They adopt a non-linear model that accurately captures the "convex-concave" behavior of practical rectifiers. * Optimization Goal: They aim to maximize the minimum harvested DC energy among all users (ensuring fairness) over a specific time duration. * Algorithm: The resulting optimization problem is "non-convex" and mathematically "thorny" because the time durations and the beamforming shapes are coupled. To solve this, they use: * Alternating Optimization: They fix the beamforming shapes to optimize the time slots, then fix the slots to optimize the shapes, repeating until convergence. * Successive Convex Approximation (SCA): They use Taylor expansions to linearize the difficult sigmoid functions into manageable convex sub-problems.

Datasets

As a theoretical communication theory paper, no external real-world datasets were used. Instead, the authors evaluated their approach using Monte Carlo simulations based on standard wireless channel models:

  • Rician Fading Channels: Used to model the signal path between the transmitter and receivers, accounting for both line-of-sight and scattered signals.
  • Standard Parameters: Simulations assumed a transmitter with 4 antennas and 30 receivers, with hardware parameters (like saturation power) pulled from existing empirical studies on energy harvesting circuits.

Key Insights and Ideas

  • The Failure of "Steady and Slow": The most important insight is that when received power is low, "pulsing" high power to users (via TDMA or time-sharing) is more efficient than a constant low-power stream. This is because it pushes the receiver's hardware into the more efficient part of its conversion curve.
  • A Unified Framework: The proposed time-division approach is a "super-strategy." It can naturally collapse into pure multi-beam transmission (if users are in the concave/high-power regime) or pure TDMA (if users are in the very low-power regime) depending on what is mathematically optimal for the current environment.
  • Exploiting High-Rank Matrices: Unlike traditional data beamforming which often uses "rank-1" (single direction) beams, energy beamforming here benefits from "high-rank" matrices that can balance energy across multiple users simultaneously within a single time slot.

Key Results

  • Performance Gain: In the low-power regime (e.g., transmit power around 38 dBm), the proposed design achieved approximately 7% higher harvested DC energy compared to conventional multi-beam methods.
  • Robustness to Antenna Scaling: As the number of antennas increases, the "array gain" naturally pushes users out of the inefficient convex regime and into the concave regime. At this point, the proposed method automatically converges to the performance of multi-beam transmission, proving it is robust across different hardware configurations.
  • Superiority over TDMA: The design consistently outperformed pure TDMA-based energy transfer, proving that a hybrid approach (sharing beams while also sharing time) is the most efficient way to handle multiuser non-linear harvesting.

Limitations and Future Work

  • CSI Requirements: The model assumes the transmitter has perfect Channel State Information (CSI). In practice, obtaining CSI for 30+ passive energy receivers is difficult and consumes energy itself.
  • Hardware Complexity: Implementing rapidly switching beamforming matrices requires sophisticated RF chain synchronization.
  • Future Directions: The authors suggest that future work could explore dynamic environments where users move, or scenarios where the transmitter must balance wireless power transfer with wireless information transmission (SWIPT).

Summary updated

Abstract

Energy beamforming has emerged as a promising technique for enhancing the energy transfer efficiency of radio-frequency (RF) enabled wireless power transfer (WPT). However, the performance of conventional energy beamforming may seriously degrade due to the non-linear RF to direct current (DC) conversion at energy receivers (ERs). To tackle this issue, this letter proposes a new time-division energy beamforming scheme, in which different energy beamforming matrices (of high ranks in general) are time shared to exploit the “convex-concave” shape of the RF-DC power relation at ERs. In particular, we maximize the minimum harvested DC energy among all ERs, by jointly optimizing the energy beamforming matrices and the corresponding time allocation subject to a given charging time duration (e.g., a second). To solve the formulated non-convex min-DC-energy maximization problem, we propose an efficient solution by using the techniques of alternating optimization and successive convex approximation (SCA). Numerical results show that the proposed time-division energy beamforming design indeed outperforms the conventional multi-beam and time-division-multiple-access (TDMA)-based energy transmissions.
BibTeX
@article{ma2021time,
  author = {Ganggang Ma and Jie Xu and Ya-Feng Liu and Mohammad R. Vedady Moghadam},
  title = {{Time-Division Energy Beamforming for Multiuser Wireless Power Transfer With Non-Linear Energy Harvesting}},
  journal = {IEEE Wireless Communications Letters},
  year = {2021},
  url = {https://openreview.net/forum?id=_ftMhNc8Mxj},
}

Loading related content...

Figure