Sensing-Assisted Robust SWIPT for Mobile Energy Harvesting Receivers in Networked ISAC Systems

Yiming Xu, Dongfang Xu, Shenghui Song

TL;DR

This paper introduces a two-phase framework for networked Integrated Sensing and Communication (ISAC) systems to power mobile IoT devices efficiently. By dedicating a small portion of each transmission frame to active sensing, the system "re-localizes" mobile receivers, allowing for highly accurate beamforming that significantly reduces power waste compared to traditional static methods.

What Problem Does This Paper Address?

Simultaneous Wireless Information and Power Transfer (SWIPT) is a key technology for the "battery-less" IoT future, where devices harvest energy from the same radio signals used for data. However, most existing research assumes the Energy Harvesting Receiver (EHR) is stationary.

In reality, IoT devices move. This movement, combined with initial estimation errors, creates location uncertainty. If a network tries to beam energy to where it thinks a device is, but the device has moved, the energy beam misses its target. To compensate, traditional "robust" designs often blast more power over a wider area, which is incredibly inefficient. The authors tackle this by asking: Can we use the sensing capabilities of an ISAC network to find the device first, then transfer power more precisely?

Methods

The authors propose a two-phase, time-splitting design managed by a two-layer optimization framework:

  1. Sensing Phase: Multiple Access Points (APs) collaborate to perform "echo-based" sensing. One AP acts as a receiver to catch reflections from the EHR, while others transmit. The goal here is to minimize the Cramér-Rao Bound (CRB)—a mathematical benchmark for the lowest possible sensing error—while still providing data services to regular communication users (CUs).
  2. WPT Phase: Using the refined location from Phase 1, the system performs Wireless Power Transfer (WPT). Because the location is now much more certain, the beamforming can be "sharper" and more efficient.

The Optimization Framework: * Outer Layer: Uses a line search to find the optimal time-splitting ratio (how much of the frame to spend sensing vs. powering). * Inner Layer: Solves two complex sub-problems. To handle the "messy" math of non-linear energy harvesting and binary node selection (choosing which AP senses), they employ a suite of advanced signal processing tools: Semidefinite Relaxation (SDR), Successive Convex Approximation (SCA), and the S-procedure for robustness.

Datasets

As is standard in theoretical wireless communications research, the paper relies on numerical simulations rather than a static dataset.

  • Setup: A network of 3 APs (each with 6 antennas), 5 Communication Users, and 1 mobile EHR.
  • Channel Models: They use Rayleigh fading models with specific path-loss exponents (3 for AP-to-user links, 2 for AP-to-EHR links) to simulate a realistic industrial or warehouse environment.
  • Benchmarks: The proposed scheme is compared against four baselines, including traditional robust designs without sensing and "non-robust" designs that ignore location error entirely.

Key Insights and Ideas

  • Sensing as an Investment: The "aha moment" is that spending time and power on sensing isn't "wasted" overhead. It is an investment that pays dividends in the second phase by allowing for much higher power transfer efficiency.
  • The "Goldilocks" Ratio: There is an optimal time-splitting ratio. If you sense for too long, you run out of time to actually transfer power. If you sense for too short a time, your location data is too noisy to be useful.
  • Macro-diversity: By using a networked (collaborative) approach, the system benefits from seeing the EHR from multiple angles, which drastically improves sensing accuracy compared to a single AP trying to do it all.

Key Results

  • Power Efficiency: The proposed sensing-assisted scheme consistently consumes less average transmit power than all baseline schemes. As location uncertainty increases, the "power savings" gap between this method and traditional robust designs widens significantly.
  • Robustness: The system maintains a much lower infeasibility rate. In scenarios where traditional designs fail to meet the device's minimum energy requirements (because the location error is too high), the sensing-assisted approach still succeeds.
  • Fast Convergence: Despite the complexity of the two-layer nested optimization, the algorithms converge in fewer than 10 iterations, suggesting they could be practical for real-world implementation.

Limitations and Future Work

  • Single EHR Focus: The current model focuses on a single mobile EHR. In a dense IoT environment (like a smart factory), the system would need to manage multiple mobile devices simultaneously, which complicates the sensing and interference management.
  • Computational Overhead: While the algorithm converges quickly, the nested-loop structure and SDR/SCA techniques are computationally intensive. Future work might look into deep learning-based "unfolding" to speed up these calculations for real-time hardware.
  • Synchronization: The paper assumes perfect synchronization across the networked APs (a Cloud-RAN architecture), which can be challenging to maintain in very large-scale deployments.

Summary updated

Abstract

Simultaneous wireless information and power transfer (SWIPT) has been proposed to offer communication services and transfer power to the energy harvesting receiver (EHR) concurrently. However, existing works mainly focused on static EHRs, without considering the location uncertainty caused by the movement of EHRs and location estimation errors. To tackle this issue, this paper considers the sensing-assisted SWIPT design in a networked integrated sensing and communication (ISAC) system in the presence of location uncertainty. A two-phase robust design is proposed to reduce the location uncertainty and improve the power transfer efficiency. In particular, each time frame is divided into two phases, i.e., sensing and WPT phases, via time-splitting. The sensing phase performs collaborative sensing to localize the EHR, whose results are then utilized in the WPT phase for efficient WPT. To minimize the power consumption with given communication and power transfer requirements, a two-layer optimization framework is proposed to jointly optimize the time-splitting ratio, coordinated beamforming policy, and sensing node selection. Simulation results validate the effectiveness of the proposed design and demonstrate the existence of an optimal time-splitting ratio for given location uncertainty.
BibTeX
@article{xu2025sensing,
  author = {Yiming Xu and Dongfang Xu and Shenghui Song},
  title = {{Sensing-Assisted Robust SWIPT for Mobile Energy Harvesting Receivers in Networked ISAC Systems}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2025},
  url = {https://openreview.net/forum?id=rcF7dVCbfK},
}

Loading related content...

Figure