Structured OMP for IRS-aided mmWave Channel Estimation by Exploiting Angular Spread

An emerging technology namely the intelligent reflecting surface (IRS) can be deployed in millimeter wave (mmWave) communication to overcome the large propagation loss and huge power consumption induced by the short wavelength. However, the large number of passive IRS elements without signal processing abilities induces high pilot overhead for channel estimation (CE). In this paper, the angular spread feature or also referred to the cluster feature is exploited to formulate the CE problem as a structured sparse signal recovery problem. Then, structured orthogonal matching pursuit (S-OMP) algorithm is proposed to efficiently solve the problem by utilizing the structure of the channel matrix in row and column simultaneously. Simulation results demonstrate that S-OMP reduces more than 40% pilot overhead while maintaining the same performance compared with the existing methods.