IRS-Assisted MISO with Finite-Alphabet Inputs Using Two-Timescale CSI

This paper analyzes the ergodic mutual information (EMI) of an intelligent reflecting surface (IRS)-aided multiple-input single-output system with finite-alphabet inputs relying on two- timescale channel state information. For the sake of unveiling important system design insights, asymptotic analyses are performed on the EMI in the regime of high signal-to-noise ratio (SNR). It is found that the EMI converges to some constant in the high-SNR regime and the rate of convergence is determined by the diversity order and the array gain. On this basis, two efficient algorithms are proposed to design the phase shifts of the IRS in order to improve the array gain as well as the EMI. All the analytical results are verified through computer simulations.