Simple Semi-Grant-Free Transmission Strategies Assisted by Non-Orthogonal Multiple Access

Zhiguo Ding, Robert Schober, Pingzhi Fan, H. Vincent Poor

TL;DR

This paper introduces Semi-Grant-Free (SGF) transmission, a hybrid protocol that allows opportunistic "grant-free" users to share a channel already reserved for a high-priority "grant-based" user. By leveraging Non-Orthogonal Multiple Access (NOMA), the authors demonstrate how to maximize spectral efficiency and connectivity without sacrificing the reliability of primary users.

What Problem Does This Paper Address?

As we move toward massive IoT (mMTC) and ultra-reliable low-latency communications (URLLC), traditional "grant-based" protocols—where a device must handshake with a base station before sending data—become a bottleneck due to massive signaling overhead.

The alternative, "grant-free" transmission, allows devices to transmit whenever they have data. However, this leads to a "wild west" scenario: if too many devices transmit simultaneously on the same channel, collisions occur, and the base station’s multi-user detection (MUD) fails. This paper seeks a "Goldilocks" solution: a way to support massive connectivity with low overhead while maintaining strict control over interference to ensure high-priority users aren't drowned out.

Methods

The authors propose NOMA-assisted SGF. In this model, one high-priority user ($U_0$) is granted a dedicated resource block. Other opportunistic users ($U_m$) are allowed to "piggyback" on this same block using NOMA principles.

To prevent the opportunistic users from crashing the system, the paper introduces two contention control mechanisms:

  1. Open-Loop Control: The base station broadcasts a channel quality threshold ($\tau$). Users only transmit if their channel gain satisfies a specific condition relative to $\tau$. This requires zero individual handshaking.
  2. Distributed Contention Control: A timer-backoff mechanism where users with the "best" (or "worst," depending on the strategy) channel conditions transmit a beacon first, effectively self-selecting a fixed number of admitted users.

The paper analyzes two distinct scenarios based on Successive Interference Cancellation (SIC) decoding orders:

  • Type I SGF: The high-priority user is decoded first. This is ideal when the priority user has a strong signal (near the base station) and the grant-free users are at the cell edge.
  • Type II SGF: The high-priority user is decoded last. This is best when the priority user is at the cell edge and needs the grant-free users to be "strong" NOMA partners to facilitate successful decoding.

Datasets

This is a theoretical and analytical paper. No external real-world datasets were used. Instead, the authors validated their mathematical derivations using Monte Carlo computer simulations based on standard independent and identically distributed (i.i.d.) quasi-static Rayleigh fading channel models.

Key Insights and Ideas

  • The "Anchor" Concept: By using a grant-based user as an "anchor," the base station can maintain a baseline of spectral efficiency while treating the remaining capacity as an opportunistic pool for grant-free devices.
  • Decoding Order is Context-Dependent: There is no "one size fits all" for NOMA. If your primary user is weak (cell-edge), you actually want strong grant-free users to transmit so they can be decoded and removed first, leaving a cleaner signal for the primary user.
  • Contention Control as a Safety Valve: The distributed contention control mechanism is particularly powerful because it ensures that even if the number of potential grant-free users ($M$) goes to infinity, the outage probability for the primary user remains stable.

Key Results

  • Performance Parity: Analytical and simulation results show that SGF can allow multiple grant-free users to transmit while the primary user experiences nearly the same outage probability as if they were alone on the channel.
  • Asymptotic Stability: In the distributed control case, the "extra cost" (interference) of admitting grant-free users approaches zero as the number of available grant-free users increases, because the system can always pick the most "compatible" users for NOMA.
  • Validation: The paper provides closed-form expressions for outage probabilities that perfectly match simulation results, providing a robust framework for network engineers to set thresholds ($\tau$) in real-world deployments.

Limitations and Future Work

  • Perfect CSI Assumption: The current model assumes users have perfect Channel State Information (CSI). In reality, channel estimation error is a significant factor in NOMA performance.
  • Perfect SIC: The analysis assumes that Successive Interference Cancellation is perfect. Residual interference from imperfect cancellation would likely introduce error floors not fully explored here.
  • Future Directions: The authors suggest that investigating the impact of imperfect CSI and more complex multi-antenna (MIMO) scenarios are the next logical steps for refining SGF strategies.

Summary updated

BibTeX
@article{ding2019simple,
  author = {Zhiguo Ding and Robert Schober and Pingzhi Fan and H. Vincent Poor},
  title = {{Simple Semi-Grant-Free Transmission Strategies Assisted by Non-Orthogonal Multiple Access}},
  journal = {IEEE Trans. Communications},
  year = {2019},
  doi = {10.1109/tcomm.2019.2903443},
}
Figure