Optimal Placement of Reconfigurable Intelligent Surface for Millimeter-Wave Indoor Communication

Recently, millimeter-wave or terahertz communications have been actively studied because of their potential for large bandwidth utilization. However, due to shorter wavelengths, the line-of-sight (LoS) path is the main contributor to signal propagation. For indoor communication, however, LoS paths are hard to exist due to various obstacles such as pillars, walls, and furniture. To mitigate such signal blockage, reconfigurable intelligent surface (RIS) is introduced in this paper. We focus on the RIS placement optimization to maximize the coverage area improved by the RIS. The proposed supervised learning approach achieves a coverage area close to that of an exhaustive search with low computational complexity and also outperforms the conventional schemes attained by the decision tree algorithm and the randomly deployed RIS.