Research Article
Deep Leaning Aided NOMA Combining Different NOMA Schemes
@INPROCEEDINGS{10.1007/978-3-030-93398-2_69, author={Qiuyi Sui and Shaochuan Wu and Haoran Zhang}, title={Deep Leaning Aided NOMA Combining Different NOMA Schemes}, proceedings={Wireless and Satellite Systems. 12th EAI International Conference, WiSATS 2021, Virtual Event, China, July 31 -- August 2, 2021, Proceedings}, proceedings_a={WISATS}, year={2022}, month={1}, keywords={NOMA Deep learning SCMA}, doi={10.1007/978-3-030-93398-2_69} }- Qiuyi Sui
Shaochuan Wu
Haoran Zhang
Year: 2022
Deep Leaning Aided NOMA Combining Different NOMA Schemes
WISATS
Springer
DOI: 10.1007/978-3-030-93398-2_69
Abstract
Non-orthogonal multiple access (NOMA) is a promising technique for future wireless communication. Compared with orthogonal multiple access (OMA), it provides high spectral efficiency and the ability to support a large number of users. A novel DNN-NOMA scheme is proposed in this paper. Both the encoder and decoder of it are composed of deep neural networks (DNN). This DNN-NOMA scheme combines power-domain NOMA, sparse code multiple access (SCMA) and multi-user shared access (MUSA), which means all of these NOMA schemes can be regarded as a special case of it. DNN-based encoders and decoders are able to generate appropriate codebooks and decode received signals automatically. The orthogonal resources (OR) used by each user are automatically determined in the process of optimizing network parameters. The simulation results prove that the BER of this novel DNN-NOMA scheme is lower than other NOMA schemes. Moreover, because there is no need to design the codebook manually, it can easily adapt to different numbers of users and ORs.
