Impact of Spatial Correlation in Element Selection and Grouping for Irregular BD-IRS

Intelligent reflecting surface (IRS) typically employs a diagonal scattering matrix to independently adjust the phase of each reflecting element, whereas beyond-diagonal IRS (BD-IRS) breaks this limitation by interconnecting IRS elements, thereby achieving higher degrees of freedom. Based on this concept, group-connected IRS groups the elements to balance circuit complexity and system performance. On the other hand, irregular IRS utilizes non-uniform element placement to more effectively exploit the varying channel conditions among IRS elements. In this letter, we investigate the impact of spatial correlation on element selection and grouping in an irregular group-connected IRS system. The study reveals that when internal correlation is reduced and external correlation is increased, the transmission performance at the receiver is significantly enhanced. Based on this observation, a low-complexity greedy-based algorithm is proposed. Simulation results validate the advantages of the proposed method in improving the received signal power and offer new insights for element selection and grouping strategies.