Energy Harvesting Noncoherent Cooperative Communications

Peng Liu, Saeed Gazor, Il-Min Kim, Dong In Kim

Abstract

Simultaneous wireless information and power transfer (SWIPT) is a key technology for self-sustaining relay networks in the Internet of Things (IoT). However, traditional coherent SWIPT systems are burdened by the high energy cost of channel state information (CSI) estimation. This paper proposes a noncoherent SWIPT framework that eliminates CSI requirements by employing M-ary Frequency-Shift Keying (FSK) and Differential Phase-Shift Keying (DPSK). We develop practical, near-optimal detectors and analyze the performance tradeoffs between energy harvesting and information decoding. Our results demonstrate that noncoherent M-FSK significantly outperforms DPSK in energy-constrained environments, providing a more efficient and hardware-friendly solution for low-power relaying.

Introduction

In the evolving landscape of the Internet of Things (IoT) and remote sensing, energy harvesting (EH) has become a critical requirement for long-term autonomous operation. SWIPT technology enables a relay node to harvest energy from a source's radio frequency (RF) signal and subsequently use that energy to forward data to a destination. While promising, most existing SWIPT research relies on coherent communication, which necessitates the receiver's knowledge of the wireless channel state (CSI).

The process of obtaining CSI—through periodic training symbols and complex calculations—consumes a significant portion of the very energy the relay is trying to harvest. This "CSI tax" increases hardware complexity and reduces the energy available for data transmission. This paper investigates whether a noncoherent approach can bypass these limitations, offering a simpler and more energy-efficient alternative.

Proposed Noncoherent Framework

We propose a unified noncoherent SWIPT framework supporting two primary relaying protocols: Power-Splitting (PS) and Time-Switching (TS). These protocols define how the relay manages the incoming RF signal for its dual purposes of energy harvesting and information decoding.

Comparison of Power-Splitting (PS) and Time-Switching (TS) protocols for noncoherent decode-and-forward (NcDF) relaying.
  • Power-Splitting (PS-NcDF): The relay divides the received signal power into two streams. One stream is directed to the energy harvester, while the other is used for noncoherent information decoding.
  • Time-Switching (TS-NcDF): The relay alternates its operation over time. A specific portion of the time block is dedicated solely to energy harvesting, and the remainder is used for information transmission.

Methods and Detection

The framework utilizes noncoherent modulations, specifically M-ary Frequency-Shift Keying (FSK) and Differential Phase-Shift Keying (DPSK). A central challenge in noncoherent systems is the complexity of the Maximum-Likelihood Detectors (MLDs), which often involve computationally intensive integrals.

To address this, we derived exact MLDs and developed closed-form approximate MLDs using Taylor series expansions and Gaussian-Legendre quadrature. These "shortcuts" enable near-optimal detection performance with significantly reduced processing power, making them suitable for energy-constrained IoT devices. Furthermore, we conducted a rigorous transition probability analysis to understand how errors propagate across the two-hop relay channel.

Key Insights and Results

Our analysis reveals several critical findings for the design of self-powered relay systems:

  1. The CSI Tax Refund: By eliminating the need for CSI, noncoherent systems effectively "refund" the energy typically spent on channel estimation, allowing it to be used for actual data forwarding.
  2. The Harvesting Tradeoff: There is a delicate balance in setting the harvesting parameters (the time ratio $\alpha$ for TS or power ratio $\rho$ for PS). Harvesting too much energy leads to decoding failures at the relay, while harvesting too little leaves the relay with insufficient power to reach the destination. We proved that the optimal values for these parameters are strictly between 0 and 1.
  3. M-FSK vs. DPSK: While DPSK is common in traditional systems, M-FSK proves superior for energy-harvesting relays at higher modulation orders ($M \ge 8$). The orthogonality of FSK makes it more resilient and energy-efficient in this power-constrained, noncoherent context.

Experimental results using Monte Carlo simulations confirm that our approximate MLDs perform almost identically to exact MLDs across the entire SNR range. Notably, for a target Symbol Error Rate (SER) of $10^{-2}$, M-FSK provides a performance gain of approximately 4 dB at $M=8$ and 8 dB at $M=16$ compared to M-DPSK.

Conclusion

This work demonstrates that noncoherent communication is a highly effective strategy for SWIPT-enabled relay networks. By removing the burden of CSI estimation and leveraging the strengths of M-FSK, we can build simpler, more reliable, and more energy-efficient IoT systems. Future research will extend this framework to multi-antenna (MIMO) configurations and explore specialized noncoherent space-time codes to further enhance system reliability.

Summary updated

BibTeX
@article{liu2015energy,
  author = {Peng Liu and Saeed Gazor and Il-Min Kim and Dong In Kim},
  title = {{Energy Harvesting Noncoherent Cooperative Communications}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2015},
  doi = {10.1109/twc.2015.2458969},
}
Figure