Enhancing active reconfigurable intelligent surface

Abstract

<p> A Reconfigurable Intelligent Surface (RIS) panel comprises many independent Reflective Elements (REs). One possible way to implement an RIS is to use a binary passive load impedance connected to an antenna element to achieve the modulation of reflected radio waves. Each RE reflects incoming waves (incident signal) by using on/off modulation between two passive loads and adjusting its phase using a Phase Shifter (PS). However, this modulation process reduces the amplitude of the reflected output signal to less than unity. Therefore, recent RIS works have employed Reflection Amplifiers (RAs) to compensate for the losses incurred during the modulation process. However, these systems only improve the reflection coefficient for a single modulation state, resulting in suboptimal RE efficacy. Thus, this paper proposes a strategy for optimising RE by continuously activating the RA regardless of the switching load state. The performance of the proposed scheme is evaluated in two scenarios: (1) In the first scenario (Sc1), the RA only operates to compensate for high-impedance loads, and (2) in the second scenario (Sc2), the RA runs continuously regardless of the RE loads. To benchmark the performance of Sc1 and Sc2, various metrics are compared, including signal-to-noise ratio, insertion loss, noise figure, communication range, and power-added efficiency. Numerical examples are provided to demonstrate the effectiveness of the proposed scheme. It is found that the proposed system in Sc2 leads to better overall performance compared to Sc1 due to the increased gain of the RIS reflection.</p>

THz Communications: A Key Enabler for Future Cellular Networks

“…Most of the existing research on IRS largely adheres to the conventional model-driven beamforming design approach, relying heavily on precise and real-time channel information. However, obtaining accurate channel state information for IRS-aided communication systems is quite challenging due to factors such as complexity, cost, and protocol compatibility [129]. To overcome such difficulties, several studies have explored the application of neural networks for channel estimation or beamforming of IRS [130]- [134].…”