Deep Learning-based Joint Symbol Detection for NOMA
Ahmet Emir, Ferdi Kara, Hakan Kaya, None None
proceeding (Uluslararası, Hakemli) | Yazar(lar): Ahmet Emir, Ferdi Kara, Hakan Kaya, None None | Yıl: 2019 | Dergi/Kurum: 2019 27TH SIGNAL PROCESSING AND COMMUNICATIONS APPLICATIONS CONFERENCE (SIU) | İndeks: WOS.ISTP | DOI: 10.1109/siu.2019.8806600 | Anahtar Kelimeler: Non-orthogonal multiple Access (NOMA), deep learn
Non-orthogonal multiple access is one of the most important candidates for next-generation communication systems called 5G and beyond. NOMA provides superiority to multiple access techniques that has been used up to recent times in terms of outage probability and achieved throughput. However, two main drawbacks of NOMA are as follow: 1) The decay in error performance caused by inter user interference. 2) The high computational complexity at receivers because of successive interference canceler (SIC). In this study, Deep Learning (DL)-based joint symbol detection to detect symbols at users is proposed. Based on the provided simulation results we present that DL based detector can achieve the same error performance with SIC based detector. Hence, the power of DL networks for wireless communications has been revealed.