PROJECT TITLE :

STAR-RIS Integrated Nonorthogonal Multiple Access and Over-the-Air Federated Learning Framework, Analysis, and Optimization

ABSTRACT:

In this paper, nonorthogonal multiple access (NOMA) and over-the-air federated learning (AirFL) are combined into a single, unified architecture through the utilization of a single, simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS). Adjusting the decoding order of hybrid users is an important function that the STAR-RIS performs for the purpose of efficient interference mitigation and extension of omnidirectional coverage. A closed-form expression for the optimality gap, also known as the convergence upper bound, is derived in order to capture the effect that nonideal wireless channels have on AirFL. This expression is used to determine the difference between the actual loss and the optimal loss. Based on this analysis, it can be deduced that the active and passive beamforming strategies, in addition to wireless noise, have a significant impact on the performance of the learning process. In addition, it has been demonstrated beyond a reasonable doubt that the optimality gap will converge with a linear rate in the event that the learning rate decreases as the training goes on. A mixed-integer nonlinear programming (MINLP) problem is formulated by jointly designing the transmit power at users and the configuration mode of STAR-RIS in order to accelerate convergence while simultaneously satisfying quality-of-service requirements. This is done in order to ensure that the requirements are met. Next, in order to handle the decoupled nonconvex subproblems in an iterative manner, a trust-region-based successive convex approximation method and a penalty-based semidefinite relaxation approach are both proposed as potential solutions. After that, an alternating optimization algorithm is developed in order to locate a suboptimal solution for the MINLP problem that was initially presented. Extensive simulation results show that: 1) the proposed framework can efficiently support NOMA and AirFL users via concurrent uplink communications; 2) our algorithms achieve a faster convergence rate on independent and identically distributed (IID) and non-IID settings compared to the existing baselines; and 3) both the spectrum efficiency and learning performance are significantly improved with the assistance of the well-tuned STAR-RIS. These findings are supported by the following hypotheses: 1) the proposed framework can efficiently support NOMA and AirFL users