Deep learning optimization of STAR-RIS for enhanced data rate and energy efficiency in 6G wireless networks

Area of Science:

• Wireless Communications
• Intelligent Surfaces
• 6G Networks

Background:

• Traditional Reconfigurable Intelligent Surfaces (RISs) are limited to half-space coverage.
• Simultaneously Transmitting and Reflecting RISs (STAR-RISs) enable full-space coverage, offering new degrees of freedom for signal propagation control.
• STAR-RIS technology is a key enabler for future 6G wireless networks.

Purpose of the Study:

• To evaluate the performance of STAR-RIS in 6G wireless networks.
• To compare nearly passive STAR-RIS (NP-STAR) and active STAR (ASTAR) against nearly passive RIS (NP-RIS) and active RIS (ARIS) benchmarks.
• To assess achievable data rates and spectral energy efficiency (SEE) using deep learning (DL)-based optimization.

Main Methods:

• Extensive simulations were conducted to analyze system performance.
• Deep learning (DL) based optimization was employed for performance assessment.
• Performance metrics including achievable data rates and spectral energy efficiency (SEE) were evaluated across various configurations.

Main Results:

• In high-interference environments, NP-STAR configurations demonstrated superior performance over ASTAR by mitigating interference amplification.
• Active RIS implementations exhibited greater energy consumption compared to nearly passive ones, leading to reduced spectral energy efficiency (SEE).
• Deep learning approaches effectively approximated genie-aided performance bounds with sufficient training data, particularly in complex channel scenarios.

Conclusions:

• STAR-RIS technology presents significant potential for 6G wireless networks, offering full-space coverage.
• Managing base station transmit power is crucial for interference control, with active STAR-RIS excelling in optimal channel conditions.
• Addressing research gaps in real-world performance, energy-spectral efficiency trade-offs, and DL optimization is vital for practical STAR-RIS deployment.