A Flexible Design for Active Reconfigurable Intelligent Surface—A Sub-Array Architecture

The emerging reconfigurable intelligent surface (RIS) technology has nowadays been cast with great attentions. The latest research has unveiled that the severe double-fading effect significantly restricts the coverage of pure passive RIS. In this context, a novel active RIS architecture has recently been proposed, where active components are employed to magnify incident signals. Although the active RIS can effectively combat the double-fading loss, one potential drawback is the active components also bring non-negligible energy expenditure. To overcome this shortcoming, we propose a novel sub-array-based RIS construction, which partitions the entire intelligent surface into multiple sub-arrays with each sub-array being flexibly switched on/off. This new architecture entitles one to wisely deactivate the sub-arrays with poor channel fading gains to save energy. Mathematically, the joint sub-array activation and beamforming design towards power consumption and energy efficiency (EE) optimization is highly challenging due to its combinatorial nature. To resolve this challenge, we combine the group sparsity inducing method and the majorization-minimization (MM) framework to develop efficient solutions. Numerical results demonstrate that our proposed solutions can achieve nearly identical performance with the exhaustive-search methods, but with a much lower complexity. Besides, our proposed sub-array-based RIS architecture can significantly improve the power and EE performance.