Sum-rate Maximization for Intelligent Reflecting Surface Based Terahertz Communication Systems

Terahertz (THz) communication system is envisioned as a promising alternative to support ultra-high speed data transmission for future indoor application scenarios. Due to the existence of the potential obstacles, the line-of-sight communication links for indoor THz communication are not reliable. In this paper, we thus investigate the utilization of intelligent reflecting surface (IRS) to enhance the reflecting transmission of the THz communication system. Specifically, an IRS consists of a large number of reflecting elements, and the phase-shift of each reflecting element is adjustable. Based on the principle of IRS, the propagation direction of THz signals can be changed via adjusting all the phase-shifts of IRS, and then we are able to improve the sum-rate performance by selecting the optimal values of the phase-shifts. Accordingly, we first propose a local search (LS) method, which can greatly decease the complexity compared with the traditional exhaustive search method. However, the LS method suffers a certain performance loss. To this end, we then propose a cross-entropy (CE) method that is feasible to promote the sum-rate compared with the LS method. Numerical results verify the above conclusions, and also show the merit of the IRS-enhanced THz communication system.