Practical Evaluation Method for Reflected Power of Electrically Large IRS in the Near-Field Region: Synthesis of Reflected E-Field of Sub-IRS

The use of intelligent reflecting surfaces (IRSs) to enhance coverage in millimeter-wave bands has attracted significant attention. To utilize IRS in mobile networks, it is necessary to determine the reflected power of the IRS. As the electrical size of the IRS must be large to enhance the reflected power, its application in the near-field region (Fresnel region) should increase. In addition, an actual IRS is affected by unaccounted errors such as reflection phase errors, reflection amplitude errors, and mutual coupling. A practical evaluation method for electrically large IRS is required for highly accurate reflected power estimation. To address these problems, we propose dividing electrically large IRS into sub-IRSs to synthesize the reflected E-field from each sub-IRS. The reflected E-field is estimated using the measured radar cross-section (RCS) of the sub-IRSs. To compensate for the differences in the incident and reflection angles of each sub-IRS, we proposed estimating the difference using a geometrical approach. The power reflected from each sub-IRS is estimated by shifting the reflection angle in the RCS pattern. The numerical results verified that the reflection power of the electrically large IRS was estimated with high accuracy while the location of each transmitting point and observation point met the far-field condition of the sub-IRS. The measurement results in an anechoic chamber also verified that the proposed method could estimate the reflected power with high accuracy compared with the conventional physical optics (PO)-based approach. The median power difference of the proposed method and measured results was approximately 3 to 4 dB, and the correlation between them exceeded 0.77, which are superior to those of the conventional PO-based approach.