Reconfigurable Intelligent Surface Aided Constant-Envelope Wireless Power Transfer

Xiaojun Yuan, Huiyuan Yang, Xiaojun Yuan 0002, Jun Fang, Ying-Chang LiangIEEE TSP 2021 · DOI

TL;DR

This paper proposes a novel method to improve wireless power transfer (WPT) efficiency by using reconfigurable intelligent surfaces (RISs) to intelligently shape the radio frequency (RF) environment. They develop an algorithm to optimize both the transmitter's beamforming and the RIS's phase shifts to maximize the total power delivered to multiple users while ensuring each user receives a minimum power level.

What Problem Does This Paper Address?

Wireless power transfer (WPT) holds immense promise for charging devices without cables, but its efficiency plummets over longer distances. Traditional solutions like energy beamforming, which focuses RF energy towards users, can be costly due to the need for individual radio frequency (RF) chains per antenna. Constant-envelope analog beamforming offers a hardware-cost reduction by sharing an RF chain and using only phase shifters. However, even with these advancements, achieving efficient and fair power delivery to multiple users remains a challenge, especially when users have varying power requirements or are located in complex environments. This paper tackles this by introducing Reconfigurable Intelligent Surfaces (RISs) as a way to actively manipulate the RF propagation environment to boost WPT efficiency and fairness.

Methods

The core of this paper lies in its optimization framework and the algorithm developed to solve it. The authors formulate a problem to maximize the total received power at all users, subject to two key constraints:

  1. Constant-envelope constraint: The signals transmitted by the base station (BS) must have a constant amplitude.
  2. Minimum received power constraint: Each user must receive at least a specified minimum power level.

This optimization problem is non-convex, making it difficult to solve directly. To overcome this, they propose a SPMC-SCA-ADMM algorithm, which combines several techniques:

  • Alternating Optimization: The problem is broken down into sub-problems, optimizing the transmitter's beamforming vector ($\mathbf{x}$) and the RIS phase shifts ($\mathbf{\Psi}$) iteratively.
  • Successive Convex Approximation (SCA): For the non-convex parts of the problem (specifically, the constant-envelope constraint), a lower bound is derived using Taylor expansion, transforming it into a convex problem that can be handled.
  • Alternating Direction Method of Multipliers (ADMM): This method is employed to efficiently solve the resulting convex sub-problems, particularly those involving multiple local variables and constraints.

The algorithm iteratively refines the solutions for $\mathbf{x}$ and $\mathbf{\Psi}$ until convergence.

Datasets

Not discussed in the paper. The paper relies on simulated channel models (Rician and Rayleigh fading) and system parameters for evaluation.

Key Insights and Ideas

  • RISs as Programmable Environments: The paper highlights the power of RISs not just as passive reflectors but as active components that can be "programmed" to sculpt the RF propagation environment, significantly enhancing WPT.
  • Joint Optimization is Crucial: Simply optimizing the transmitter or the RIS in isolation is insufficient. The paper demonstrates that jointly optimizing both the transmitter's beamforming and the RIS's phase shifts yields superior results.
  • Balancing Efficiency and Fairness: The formulation explicitly addresses the trade-off between maximizing total power (efficiency) and ensuring each user gets a minimum amount (fairness), which is critical for practical WPT systems.
  • Low-Complexity Algorithm: The proposed SPMC-SCA-ADMM algorithm is designed to be computationally efficient, making it practical for real-world deployment, especially in scenarios with many users and RIS elements.

Key Results

The numerical results demonstrate the significant benefits of using RISs:

  • Increased Total Received Power: The presence of RISs, even when fixed (not optimized), substantially increases the total received power compared to a system without RISs.
  • Optimized RIS Yields More Power: Further improvements in total received power are achieved when the RIS phase shifts are actively optimized.
  • Impact of RIS Size: A larger number of RIS elements (N) leads to greater improvements in total received power.
  • Trade-off with Fairness: The paper shows that increasing the minimum power requirement for users (controlled by parameter $\gamma$) leads to a higher minimum received power but a decrease in the total received power, illustrating the inherent efficiency-fairness trade-off.

Limitations and Future Work

  • Channel State Information (CSI): The proposed algorithm assumes perfect knowledge of all channel state information (BS-user, BS-RIS, and RIS-user links). In practice, acquiring accurate CSI can be challenging, especially in dynamic environments.
  • Simulation-Based Evaluation: The results are based on simulations using modeled channel conditions. Real-world deployment might introduce additional complexities and performance variations.
  • Hardware Imperfections: The model assumes ideal RIS elements and constant-envelope transmission. Real hardware may exhibit non-ideal behavior, such as phase noise or amplitude variations.

Future work could explore robust algorithms that are less sensitive to CSI errors, investigate the impact of hardware impairments, and extend the framework to more complex scenarios like dynamic user mobility or the integration of multiple RISs with different functionalities.

Summary updated

BibTeX
@article{yuan2021reconfigurable,
  author = {Xiaojun Yuan and Huiyuan Yang and Xiaojun Yuan and Jun Fang and Ying-Chang Liang},
  title = {{Reconfigurable Intelligent Surface Aided Constant-Envelope Wireless Power Transfer}},
  journal = {IEEE TSP},
  year = {2021},
  doi = {10.1109/tsp.2021.3056906},
}
Figure