Broadcast Channels With Heterogeneous Arrival and Decoding Deadlines: Second-Order Achievability

Homa Nikbakht, Malcolm Egan, Jean-Marie Gorce, H. Vincent Poor

Abstract

In the evolution toward 5G and 6G wireless networks, Ultra-Reliable Low-Latency Communications (URLLC) demand the transmission of short packets with strict, often overlapping, deadlines. This paper investigates the second-order achievability of broadcast channels under such heterogeneous arrival and decoding constraints. By adapting Marton’s coding strategy to the finite blocklength regime and employing a three-phase transmission architecture, the authors derive rigorous performance bounds. These results demonstrate significant gains over traditional time-sharing and highlight the inaccuracies of standard normal approximations in short-packet scenarios.

Introduction

The next generation of wireless communication is defined by applications requiring extreme reliability and minimal latency, such as autonomous driving, remote surgery, and industrial automation. In these URLLC scenarios, data packets are typically short, and the assumption of infinite blocklength—central to classical information theory—no longer holds. Furthermore, high-frequency updates often lead to situations where a new message arrives before the previous one has reached its decoding deadline, creating a "heterogeneous" environment with overlapping transmission windows.

The Challenge of Heterogeneous Deadlines

Standard information-theoretic tools often treat messages as independent entities if they arrive at different times. However, when deadlines overlap, treating messages independently through time-sharing is inefficient. The core challenge lies in jointly encoding these messages to maximize reliability and data rate within very short blocklengths (often $n < 200$), where the "Normal Approximation" frequently fails to provide accurate performance guarantees.

Proposed Methodology: Three-Phase Joint Encoding

The researchers propose a novel coding scheme for a two-user Gaussian Broadcast Channel (BC) based on four technical pillars:

  1. Marton’s Coding Strategy Adaptation: The authors adapt this classic asymptotic strategy for the finite blocklength regime, allowing for "pre-coding" that manages interference between receivers.
  2. Three-Phase Transmission Architecture: The encoding process is structured based on the availability of messages:
  3. Phase 1: Only the first message ($m_1$) is transmitted.
  4. Phase 2: Both messages ($m_1$ and $m_2$) are present and jointly encoded.
  5. Phase 3: After $m_1$ reaches its deadline, only $m_2$ is transmitted.
  6. Power-Shell Codebooks: Codewords are distributed uniformly on a multi-dimensional sphere, which offers superior performance near channel capacity for finite blocklengths compared to standard Gaussian codebooks.
  7. Second-Order Analysis: The study derives achievability bounds using second-order coding theory to account for the rate penalty inherent in short-packet communications.
System model and transmission timeline for heterogeneous broadcast channels.

Key Insights

Jointness and Efficiency

The study proves that joint encoding during the overlap period (Phase 2) is significantly more efficient than treating messages as separate, time-shared streams. By framing the temporal scheduling problem as a set of parallel channels, the authors leverage multi-user information theory to optimize the transmission.

The Normal Approximation Gap

A critical finding is that the widely used Normal Approximation is unreliable for the extremely short blocklengths typical of URLLC. The proposed second-order bounds are tighter and provide more accurate reliability guarantees for channel uses under 200.

Correlation Control

The Marton-based scheme utilizes a correlation parameter ($\rho$) as a tuning mechanism. This allows the system to balance data rates between "strong" and "weak" receivers while maintaining an optimized total sum-rate.

Results and Discussion

Numerical evaluations of the derived mathematical bounds reveal several key outcomes:

  • Superiority over Time-Sharing: The proposed joint encoding scheme shows substantial performance gains in broadcast scenarios compared to traditional methods.
  • Accuracy at Short Blocklengths: For point-to-point links with $n < 200$, the authors' achievability bound is more precise than the standard approximations proposed in previous literature.
  • Asymptotic Consistency: As the blocklength $n$ increases, the results converge with Marton’s classic inner bound, successfully bridging the gap between finite-length practical constraints and asymptotic theory.

Conclusion and Future Work

This research provides a robust framework for handling overlapping deadlines in broadcast channels, a vital component for future URLLC systems. While the current study focuses on a two-user Gaussian model, future work could extend these derivations to $K$-user settings and explore non-Gaussian channels with complex fading. Additionally, investigating non-causal "look-ahead" scenarios where the transmitter has partial knowledge of future message arrivals could further enhance system performance.

Summary updated

BibTeX
@article{nikbakht2025broadcast,
  author = {Homa Nikbakht and Malcolm Egan and Jean-Marie Gorce and H. Vincent Poor},
  title = {{Broadcast Channels With Heterogeneous Arrival and Decoding Deadlines: Second-Order Achievability}},
  journal = {IEEE Trans. Information Theory},
  year = {2025},
  doi = {10.1109/tit.2025.3532649},
}
Figure