Enhancing Transmission of STAR-RIS-Aided Spectrum Sharing CF-CR IoT System With Element Selection Under Insufficient Power Supply at RIS

This article studies simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS)-aided spectrum sharing cell-free (CF) and cognitive radio (CR) combined the Internet of Things (IoT) system and focuses on enhancing the secondary user's (SU's) achievable rate (AR) while guaranteeing the primary user's lowest AR requirement. Different from the existing studies, we consider a special condition of limited power supply at RIS in this work, in which only a part of the electromagnetic (EM) elements in RIS can function properly due to lack of energy. To this end, we formulate a worst-case SU's AR optimization problem under both perfect channel state information (CSI) and imperfect CSI conditions. To tackle these two complicated nonconvex problems, alternating optimization (AO) framework is proposed to jointly optimize the beamformer at primary and secondary transmitters, transmitting/reflecting phase shift as well as EM element selection at STAR-RIS with provable convergence. In particular, semidefinite relaxation (SDR) + successive convex approximation + penalty-convex concave procedure (PCCP) combined algorithm and SDR+PCCP+Dinkelbach combined algorithm are proposed to solve the highly coupled EM element selection and phase shift at RIS under perfect and imperfect CSI cases, respectively. Numerical results verify that given insufficient power supply and RIS, the proposed scheme significantly improves AR of SUs and outperforms the benchmark schemes of traditional reflecting RIS-aided or no RIS-aided case as well as random EM element selection strategies.