Abstract: Energy efficiency (EE) has become a major issue in satellite communications. In this paper, we study the beamforming design strategy to maximize the EE of rate-splitting multiple access (RSMA) for multibeam satellite communications by considering imperfect channel state information at the transmitter (CSIT). We propose an expectation-based robust beamforming algorithm against the imperfect CSIT scenario. By combining successive convex approximation (SCA) with penalty function transformation, the nonconvex EE maximization problem is solved iteratively. Simulation results demonstrate the superiority of RSMA over traditional space division multiple access (SDMA) and show that our proposed algorithm achieves better EE performance than conventional benchmarks.
Introduction
As satellite networks become integral to 5G and beyond, they face two major hurdles: managing interference between multiple beams and minimizing the massive power consumption of geostationary (GEO) satellites. While previous research focused on Spectral Efficiency (SE)—maximizing data speed—this paper addresses Energy Efficiency (EE), defined as the ratio of the data rate to total power consumption.
Real-world satellite links suffer from imperfect Channel State Information at the Transmitter (CSIT) due to feedback delays and atmospheric changes like rain fading. Most existing models assume perfect CSIT, which is unrealistic for satellite links. This study proposes a robust beamforming design to maximize EE under these realistic constraints using Rate-Splitting Multiple Access (RSMA).
Methodology
The authors propose an expectation-based robust beamforming algorithm. Unlike traditional Space-Division Multiple Access (SDMA), which treats interference as pure noise, RSMA splits user messages into "common" and "private" parts. The common part is decoded by all users to mitigate inter-group interference, while private parts carry specific data.
The system model assumes a Ka-band GEO satellite using a Single Feed Per Beam (SFPB) architecture. To handle imperfect CSIT, the model assumes known channel amplitude but only statistical knowledge of the phase. The resulting EE maximization problem is highly non-convex and fractional. To solve it, the authors employ:
- Successive Convex Approximation (SCA): To turn the complex problem into a series of simpler, convex ones.
- Penalty Function Transformation: To handle rank-one constraints, ensuring the mathematical solution translates into a physically realizable beamforming vector.
- Dinkelbach-like Iteration: To handle the fractional nature of the Energy Efficiency objective function.
Simulation Environment
The proposed methods were evaluated in a standardized simulation environment based on established satellite channel models:
- Orbit: Geostationary (GEO) at 35,786 km.
- Frequency: Ka-band (20 GHz).
- Channel Effects: Log-normal distribution for rain fading and Gaussian distribution for phase errors.
- Benchmark: Results were averaged over 500 random channel realizations to ensure statistical significance.
Key Results and Analysis
The simulation results confirm that RSMA is a robustness powerhouse. As shown in Figure 1, RSMA consistently achieves higher bits-per-Joule than SDMA across all tested Signal-to-Noise Ratio (SNR) levels. A critical insight is the trade-off between speed and efficiency: maximizing for Weighted Sum Rate (WSR) leads to a significant drop in EE after an SNR threshold of approximately 18 dB, as pushing for higher rates requires disproportionate power.
Furthermore, the proposed algorithm demonstrates superior resilience to phase errors. Figure 2 illustrates that while EE decreases as phase error variance ($\delta^2$) increases, the performance gap between RSMA and SDMA widens, proving RSMA's superior robustness. The proposed "Algorithm 1" also outperforms the conventional D-WMMSE (Weighted Minimum Mean Square Error) benchmark, maintaining higher EE even under significant channel uncertainty.
Conclusion
This research demonstrates that RSMA provides a more energy-efficient and robust framework for multibeam satellite communications compared to traditional SDMA. While the current study focuses on GEO satellites and single-antenna users, future work could extend these findings to Low Earth Orbit (LEO) constellations with dynamic Doppler shifts and multi-antenna user terminals. Additionally, addressing the computational complexity of Successive Interference Cancellation (SIC) at the user terminal remains a key area for future optimization.
Summary updated