Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

End-to-End Design of Polar Coded Integrated Data and Energy Networking

Jie Hu, Jingwen Cui, Luping Xiang, Kun Yang

TL;DR

The paper presents an end-to-end polar-coded IDEN system where all transmitter and receiver components are replaced by neural networks and trained jointly as an AutoEncoder to optimize data delivery and energy harvesting. A BNN-aided polar encoder dynamically selects frozen bit positions, while a hyper-RNN decoder provides iteration-aware decoding through a hypernetwork that generates main-network weights for each decoding step. The approach integrates a learnable modulation mapper (AE-Mapper), a nonlinear energy harvesting model, and an adaptive power splitter to balance WIT and WET within a unified loss, showing improvements over conventional BP-based designs and competitive performance relative to SCL-based systems. These results highlight the potential of joint learning-based design in IDEN systems, offering enhanced adaptability and energy-aware modulation at the cost of higher computational complexity. The work advances practical IDEN deployment by enabling end-to-end optimization that accounts for nonlinear EH behavior and dynamic channel conditions, with significant implications for energy-constrained IoT networks.

Abstract

In order to transmit data and transfer energy to the low-power Internet of Things (IoT) devices, integrated data and energy networking (IDEN) system may be harnessed. In this context, we propose a bitwise end-to-end design for polar coded IDEN systems, where the conventional encoding/decoding, modulation/demodulation, and energy harvesting (EH) modules are replaced by the neural networks (NNs). In this way, the entire system can be treated as an AutoEncoder (AE) and trained in an end-to-end manner. Hence achieving global optimization. Additionally, we improve the common NN-based belief propagation (BP) decoder by adding an extra hypernetwork, which generates the corresponding NN weights for the main network under different number of iterations, thus the adaptability of the receiver architecture can be further enhanced. Our numerical results demonstrate that our BP-based end-to-end design is superior to conventional BP-based counterparts in terms of both the BER and power transfer, but it is inferior to the successive cancellation list (SCL)-based conventional IDEN system, which may be due to the inherent performance gap between the BP and SCL decoders.

End-to-End Design of Polar Coded Integrated Data and Energy Networking

TL;DR

The paper presents an end-to-end polar-coded IDEN system where all transmitter and receiver components are replaced by neural networks and trained jointly as an AutoEncoder to optimize data delivery and energy harvesting. A BNN-aided polar encoder dynamically selects frozen bit positions, while a hyper-RNN decoder provides iteration-aware decoding through a hypernetwork that generates main-network weights for each decoding step. The approach integrates a learnable modulation mapper (AE-Mapper), a nonlinear energy harvesting model, and an adaptive power splitter to balance WIT and WET within a unified loss, showing improvements over conventional BP-based designs and competitive performance relative to SCL-based systems. These results highlight the potential of joint learning-based design in IDEN systems, offering enhanced adaptability and energy-aware modulation at the cost of higher computational complexity. The work advances practical IDEN deployment by enabling end-to-end optimization that accounts for nonlinear EH behavior and dynamic channel conditions, with significant implications for energy-constrained IoT networks.

Abstract

In order to transmit data and transfer energy to the low-power Internet of Things (IoT) devices, integrated data and energy networking (IDEN) system may be harnessed. In this context, we propose a bitwise end-to-end design for polar coded IDEN systems, where the conventional encoding/decoding, modulation/demodulation, and energy harvesting (EH) modules are replaced by the neural networks (NNs). In this way, the entire system can be treated as an AutoEncoder (AE) and trained in an end-to-end manner. Hence achieving global optimization. Additionally, we improve the common NN-based belief propagation (BP) decoder by adding an extra hypernetwork, which generates the corresponding NN weights for the main network under different number of iterations, thus the adaptability of the receiver architecture can be further enhanced. Our numerical results demonstrate that our BP-based end-to-end design is superior to conventional BP-based counterparts in terms of both the BER and power transfer, but it is inferior to the successive cancellation list (SCL)-based conventional IDEN system, which may be due to the inherent performance gap between the BP and SCL decoders.
Paper Structure (26 sections, 17 equations, 12 figures, 3 tables)

This paper contains 26 sections, 17 equations, 12 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Related work
  4. New Contributions
  5. System Model
  6. Transmitter
  7. BNN-aided Polar Encoder
  8. AE-Mapper
  9. Receiver
  10. Power Splitter of Fig.\ref{['fig.structure']}
  11. AE-Demapper of Fig.\ref{['fig.structure']}
  12. EH of Fig.\ref{['fig.structure']}
  13. Hyper-RNN Decoder of Fig.\ref{['fig.BP']}
  14. End-to-End Design
  15. Optimization Problem
  16. ...and 11 more sections

Figures (12)

  • Figure 1: An end-to-end polar-coded IDEN system.
  • Figure 2: The architecture of BNN.
  • Figure 3: A generic architecture of (a) AE-Mapper and (b) AE-Demapper and EH.
  • Figure 4: An example of hyper-RNN decoding with $N=4$.
  • Figure 5: Processing element of the polar code graph.
  • ...and 7 more figures