Beyond Diagonal RIS-assisted MIMO Transmission Achieves Capacity Gains with Gradient-Based Optimization

Reconfigurable Intelligent Surfaces represent a significant advance in wireless communication, offering engineers unprecedented control over how signals travel, and researchers are now exploring ways to maximise their potential. Ainna Yue Moreno-Locubiche, Josep Vidal, and colleagues at the Universitat Politècnica de Catalunya, BarcelonaTech, investigate the benefits of a new type of surface, termed ‘beyond diagonal RIS’, in Multiple-Input Multiple-Output (MIMO) systems. Their work demonstrates that this advanced surface substantially improves spectral efficiency and coverage compared to conventional designs, particularly in millimeter wave communications where signals typically travel directly between transmitter and receiver. The team achieves this improvement through a streamlined optimisation technique for controlling the surface elements, offering a practical pathway to harness the full power of reconfigurable intelligent surfaces in future wireless networks.

BD-RIS Boosts Millimeter Wave MIMO Performance

This research demonstrates the significant performance gains achievable with beyond diagonal reconfigurable intelligent surfaces (BD-RIS) in millimeter wave multiple-input multiple-output (MIMO) downlink communication systems. Researchers focused on optimizing both transmit beamforming and MIMO capacity transmission, where signals primarily travel via direct paths between transmitter and receiver. The study successfully developed a gradient-based optimization approach, offering a simpler solution compared to previously published methods. Experiments reveal that BD-RIS significantly outperforms traditional diagonal RIS in terms of spectral efficiency and coverage.

This approach allows for multi-mode MIMO transmission, a capability not fully explored in prior research. The gradient-based solver effectively optimizes the BD-RIS settings, enhancing control over the wireless propagation channel. The research tackled the mathematical complexities inherent in integrating power allocation with BD-RIS, deriving a solution based on a simplified representation of the BD-RIS MIMO capacity problem. This allows for maximizing the potential of BD-RIS in next-generation communication systems, building upon existing research demonstrating improvements in signal quality and interference reduction. The team explored the use of graph theory for designing and optimizing BD-RIS architectures, employing a combination of theoretical analysis and simulation to evaluate performance under various conditions. Results suggest the possibility of validating these findings through real-world experiments, demonstrating that BD-RIS can significantly improve the performance of wireless communication systems, particularly in scenarios where multiple data streams are transmitted simultaneously.

Beyond Diagonal RIS Boosts Millimeter Wave Performance

This research demonstrates the significant performance gains achievable with beyond diagonal reconfigurable intelligent surfaces (BD-RIS) in millimeter wave multiple-input multiple-output downlink communication systems. By optimizing the configuration of these surfaces, the team achieved substantial improvements in both spectral efficiency and coverage compared to traditional diagonal RIS architectures. The study employed a gradient-based optimization approach, notable for its reduced complexity relative to alternative methods, to determine the optimal settings for the RIS elements. The results consistently show that BD-RIS outperforms conventional RIS, particularly in scenarios where direct signal paths dominate. Through detailed simulations, the researchers established that the Adam optimization algorithm is particularly effective in configuring BD-RIS, consistently achieving stable convergence within a limited number of iterations. The team validated these findings across a realistic deployment environment, demonstrating the robustness of the approach.

Beyond-Diagonal RIS Configuration Optimisation

This work details research into beyond-diagonal reconfigurable intelligent surfaces (BD-RIS), advanced versions of reconfigurable intelligent surfaces (RIS) that offer more flexibility in controlling signal reflection. Traditional RIS reflect signals in a limited way, while BD-RIS can manipulate both the signal’s phase and amplitude across a wider range of angles, offering significant performance improvements in wireless communication. The research aims to optimize the configuration of these surfaces to maximize data rates, coverage, and energy efficiency. Researchers focused on optimizing the BD-RIS configuration to achieve these goals, utilizing matrix derivatives and optimization techniques to design algorithms for configuring the BD-RIS.

Accurate channel modeling was crucial for designing and evaluating BD-RIS systems, and the research considered various channel models, including those that account for complex signal behavior. The team also designed algorithms for optimizing the BD-RIS configuration to achieve specific performance goals, such as maximizing data rates or extending coverage. The research presents several key contributions related to BD-RIS, including developing new models and architectures for BD-RIS, taking into account the physical properties of the reflecting elements and the desired signal manipulation capabilities. The results demonstrate that BD-RIS can significantly improve the performance of wireless communication systems, particularly in shadowed regions behind obstacles, highlighting its potential to improve coverage in challenging environments.