Abstract
As 5G and 6G networks evolve, balancing the high stability requirements of grant-based (GB) users with the low-latency needs of grant-free (GF) users becomes critical. This paper investigates Semi-Grant-Free (Semi-GF) Non-Orthogonal Multiple Access (NOMA), a hybrid strategy where GF users share spectrum with GB users. We analyze a dynamic protocol that pairs users based on instantaneous received power and derive exact analytical formulas for ergodic rates. Our findings highlight a fundamental performance ceiling for GF users and demonstrate that GB users can maintain stable throughput despite GF interference, provided the pairing constraints are met.
Introduction
Traditional wireless access relies on Grant-Based (GB) protocols, which ensure stability through a request-grant handshake but introduce significant latency. Conversely, Grant-Free (GF) access allows immediate transmission, ideal for massive IoT, but suffers from collisions. Semi-GF NOMA offers a middle ground, allowing GF users to opportunistically use the same resources as GB users. However, the random deployment of users in real-world scenarios complicates the prediction of achievable data rates. This study provides a rigorous mathematical framework to evaluate these performance limits.
Methodology
The proposed system employs a Dynamic Protocol for user pairing:
- Thresholding: The Base Station (BS) broadcasts the instantaneous received power of the GB user.
- User Selection: A GF user is permitted to transmit only if its estimated received power at the BS is lower than that of the GB user. This ensures the GB user remains the "strong" user in the NOMA pair, facilitating successful Successive Interference Cancellation (SIC).
- Mathematical Modeling: The authors utilize stochastic geometry, assuming a uniform distribution of users within a circular cell. Advanced techniques, including Chebyshev-Gauss quadrature and hypergeometric functions, are used to derive closed-form expressions for the ergodic rates of both user types.
Results and Analysis
The analytical models were validated using Monte Carlo simulations, showing a near-perfect match between theoretical derivations and simulated reality.
The GF Performance Ceiling
A key finding is the existence of a "maximum" achievable rate for GF users. As shown in Figure 1, increasing the transmit SNR of GF users ($\rho_{GF}$) initially improves their ergodic rate, but eventually leads to a decline. This occurs because excessive GF power can violate the pairing threshold or degrade the SIC process, meaning more power does not always equate to higher throughput for GF devices.
GB Stability and Interference Trade-offs
In contrast, the GB user's performance scales steadily with its transmit SNR ($\rho_{GB}$), as illustrated in Figure 2. While the presence of GF users introduces interference that reduces the GB rate, the system avoids an "error floor" through proper power management. The results confirm that increasing the transmit power of GF users directly reduces the performance of GB users because the GF signal acts as noise that the GB receiver cannot fully cancel out.
Conclusion
This research provides the first closed-form expressions for ergodic rates in Semi-GF NOMA with randomly deployed users. The results underscore the importance of instantaneous power-based pairing to maintain GB stability while enabling GF access. While the current model focuses on a single-cell scenario with Rayleigh fading, it lays the groundwork for future studies into multi-cell environments and massive Semi-GF NOMA configurations.
Summary updated