Introduction
The evolution of 5G and the upcoming 6G networks necessitates support for massive Machine-Type Communications (mMTC), where thousands of IoT sensors must coexist with traditional high-bandwidth mobile users. Traditional cellular access relies on "Grant-Based" (GB) protocols, which involve a handshake process that is inefficient for the small, sporadic data packets typical of IoT devices.
To address this, "Grant-Free" (GF) transmissions allow devices to transmit without a handshake. A more advanced approach, "Semi-Grant-Free" (SGF), enables GF users to opportunistically share frequency blocks already occupied by a primary GB user. While Non-Orthogonal Multiple Access (NOMA) has been the standard for managing interference in SGF systems, it suffers from a significant limitation: the "outage floor." This paper proposes a Rate Splitting Multiple Access (RSMA) strategy to overcome this bottleneck and enhance transmission reliability.
The Challenge: The NOMA Outage Floor
In traditional NOMA-SGF systems, if the combined target data rates of the primary GB user and the opportunistic GF user exceed a specific mathematical threshold, the system hits an outage floor. Beyond this point, increasing transmit power does not improve reliability. This limitation restricts the throughput and reliability of IoT devices in dense network environments.
Proposed RSMA-SGF Scheme
The authors introduce an RSMA-SGF scheme that replaces standard NOMA with Rate Splitting at the GF user side. This approach allows the system to handle higher target rates by strategically dividing the GF user's signal.
System Architecture and Methodology
The system consists of a Base Station (BS), a primary GB user ($U_B$), and a pool of $K$ potential GF users ($U_F$). The RSMA-SGF protocol operates through four key stages:
- Interference Thresholding: To protect the primary user's Quality of Service (QoS), the BS broadcasts a maximum interference threshold.
- Distributed Contention: GF users calculate their potential transmission rates based on this threshold. They employ a back-off timer inversely proportional to their achievable rate, ensuring the user with the best channel conditions gains access.
- Rate Splitting: If the winning GF user's signal strength would violate the interference threshold, it enters RS mode. The data is split into two streams, and transmit power is divided between them.
- Successive Interference Cancellation (SIC): The BS decodes the signals in three stages: first, one GF stream; second, the primary GB user's signal; and finally, the remaining GF stream.
Key Insights
- Eradicating the Outage Floor: By splitting the GF signal, the BS can decode just enough of it to remove it as interference for the GB user, decode the GB signal, and then decode the rest of the GF data. This mathematically bypasses the outage floor inherent in NOMA.
- Full Multiuser Diversity: The scheme achieves a diversity gain equal to $K$, the number of competing GF users. Reliability increases as more devices are added to the network, as the system capitalizes on the best available channel.
- Strict Transparency: The primary GB user experiences zero degradation. From its perspective, the network operates as if it has exclusive (Orthogonal) access to the spectrum.
Results and Performance
Through Monte Carlo simulations using Rayleigh fading channels, the authors validated their theoretical models. The results demonstrate that RSMA-SGF consistently achieves lower outage probabilities than existing benchmarks like QoS-SGF and Power Control-SGF.
Notably, when target rates are high, benchmark schemes hit an outage floor, whereas RSMA-SGF continues to improve reliability with increased power. Furthermore, the proposed scheme offers significantly higher ergodic rates (overall throughput). For instance, an RSMA-SGF setup with a single GF user can outperform a Power Control-SGF setup with five contending users at high Signal-to-Noise Ratios (SNRs).
Conclusion and Future Work
RSMA-SGF represents a significant step forward for mMTC in cellular networks, offering a robust way to integrate massive IoT devices without impacting primary users. While the current model assumes perfect Channel State Information (CSI) and a single-antenna setup, future research will explore stochastic geometry for realistic network deployments and advanced scheduling strategies to ensure fairness for users with weaker channel conditions.
Summary updated