Analysis of NOMA-OFDM 5G wireless system using deep neural network

Abstract

In this work, a multiple user deep neural network-based non-orthogonal multiple access (NOMA) receiver is investigated considering channel estimation error. The decoding of the symbol in the case of the NOMA system follows the sequential order and decoding accuracy depends on the detection of the previous user. Without estimating the throughput, a deep neural network-based NOMA orthogonal frequency division multiplexing (OFDM) system is proposed to decode the symbols from the users. Firstly, the deep neural network is trained. Secondly, the data are trained and lastly, the data are tested for various users. In this work, for various values of signal to noise ratio, the performance of the deep neural network is investigated, and the bit error rate (BER) is calculated on a per subcarrier basis. The simulation results show that the deep neural network is more robust to symbol distortion due to inter-symbol information and will obtain knowledge of the channel state information using data testing.