Power Allocation and User Grouping for NOMA Downlink Systems

Non-orthogonal multiple access (NOMA) technology allows users to use the same time-frequency resource send signals, which can improve spectral efficiency and throughput. We study problems of user grouping power allocation in downlink a multi-carrier NOMA system. The sum rate is optimization goal. A step-by-step adopted. Users are grouped first then allocated. For grouping, method based on maximum channel gain difference improved prevent with similar gains from being together. allocation, deep learning algorithm used among subcarriers. Then, closed-form expressions between multiplexed derived according minimum transmission constraint. simulation results show that compared fractional transmit fixed method, improves system by about 2.2% 19%, respectively. methods we propose 10% subcarriers users.