We consider the combination of uplink code-domain non-orthogonal multiple access (NOMA) with massive multiple-input multiple-output (MIMO) and reconfigurable intelligent surfaces (RISs). We assume a setup in which the base station (BS) is capable of forming beams towards the RISs under line-of-sight conditions, and where each RIS is covering a cluster of users. In order to support multi-user transmissions within a cluster, code-domain NOMA via spreading is utilized. We investigate the optimization of the RIS weights such that a large number of users is supported. As it turns out, it is a coupled optimization problem that depends on the detection order under interference cancellation and the applied filtering at the BS. We propose to decouple those variables by using sum-rate optimized weights as the initial solution, allowing us to obtain a decoupled estimate of those variables. Then, in order to determine the final weights, the problem is relaxed into a semidefinite program that can be solved efficiently via convex optimization algorithms. Simulation results show the effectiveness of our approach in improving the detectability of the users.

翻译：我们考虑将上链接代码-非正向多功能多存取(NOMA)与大规模多投入多输出(MIMO)和可重新配置的智能表面(RIS)相结合。我们假设一个设置,使基地台(BS)能够在视觉线下对RIS形成光束,而每个RIS覆盖一组用户。为了支持一个集群内的多用户传输,我们利用了通过扩散的代码-域域域(NOMA)来进行。我们调查了RIS加权的优化,从而支持了大量用户。事实证明,这是一个同时存在的优化问题,取决于干扰取消中的探测顺序和BS的应用过滤。我们提议通过使用超速最佳重量作为初步解决方案来区分这些变量,从而使我们能够获得这些变量的分解估计值。然后,为了确定最终的重量,问题被淡化成一个半定型程序,可以通过convex优化算法有效解决。模拟结果显示我们用户的可探测性。