The Achievable Rate Performance of STAR-RIS Aided Massive MIMO Systems

The achievable rate performance of simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) aided massive multiple-input multiple-output (MIMO) systems is investigated. Specifically, the achievable user rates are derived for three operating protocols of the STAR-RIS, namely the energy-splitting (ES), mode-switching (MS), and time-switching (TS) with both unicast and multicast transmissions. This analysis is useful in evaluating the system performance under imperfectly estimated channel state information (CSI), spatially correlated fading, pilot contamination, and statistical CSI based phase-shift optimization, transmit power control, and user signal decoding. For the high signal-to-noise ratio regime, the asymptotic achievable rates are also derived, and they serve as benchmarks or upper bounds for the rate performance comparisons for systems operating under the above transmission impediments. The composite uplink channels are estimated through linear minimum mean square error estimation technique, and the phase-shift matrices at the STAR-RIS are optimized to maximize the effective average channel gains to minimize the channel estimation overhead. The base-station optimizes the transmit power based on the max-min criterion to attain a system-wide common user rate by negating the near-far effects of the downlink composite channels. Our numerical and simulation results validate our theoretical analysis and convergence of phase-shift and transmit power optimization algorithms. Our analytical and simulation results are useful in investigating the performance gains/comparisons among the ES, MS, and TS protocols for unicast and multicast transmissions to enable 360° smart coverage extensions with passive STAR-RIS aided massive MIMO.