A comprehensive dataset of RIS-based channel measurements in the 5GHz band

Reconfigurable intelligent surfaces (RIS) are a core component considered in sixth generation (6G) communications. By utilizing an RIS prototype system in the 5 GHz band, this paper provides a comprehensive dataset of S21 channel measurements of various geometric arrangements of antennas and RIS. The dataset is provided in full on IEEE DataPort. Additionally, the in-detail documentation of the measurement setup, location, devices, geometries, and algorithms is available on GitHub. Since only a few datasets with measurements of RIS prototypes are currently available, we try to fill an important gap in current research by providing this dataset. We enable reproducible research and the validation of concepts and models with our dataset. Furthermore, the provided dataset can be used as a basis for machine learning models.The dataset includes measurements taken in stationary scenarios with specular reflection angles, non-specular reflection angles, as well as, with the RIS placed on a rotary stage. The switching state of the RIS is determined by iterative algorithms, namely a greedy algorithm and a single-element switching approach and model-based with the knowledge of the exact geometry of the setup. All switching states are optimized, once to maximize the received power and once to minimize the received power. The evaluation of the data in this paper compares the overall performance for the maximization and minimization of an RIS-assisted communication link, which shows an increase in received power of up to 20 dB compared to a reference plate.

可重构智能表面（RIS）是第六代（6G）通信的核心组件之一。本文通过在5 GHz频段使用RIS原型系统，提供了全面的数据集，其中包括不同几何排列的天线和RIS的S21通道测量数据。该数据集已完整地发布在IEEE DataPort上。此外，关于测量设置、地点、设备、几何形状和算法的详细文档可在GitHub上查阅。鉴于目前仅有少数关于RIS原型测量数据的集，我们试图通过提供此数据集来填补当前研究中的一个重要空白。我们通过提供此数据集，促进了可重复研究的实现，并验证了概念和模型。此外，所提供的数据集可作为机器学习模型的基础。数据集包括在静止场景下进行的测量，包括镜面反射角度、非镜面反射角度，以及RIS放置在旋转台上的情况。RIS的切换状态由迭代算法确定，即贪婪算法、单元素切换方法和基于设置的精确几何形状的知识进行建模的方法。所有切换状态均进行了优化，一次以最大化接收功率，一次以最小化接收功率。本文中数据的评估比较了最大化与最小化RIS辅助通信链路性能的整体表现，与参考板相比，接收功率提高了高达20 dB。