Characterizing the DNN Impact on Multiuser PD-NOMA System Based Channel Estimation and Power Allocation

This Paper demonstrates how the channel estimation based Deep Learning (DL) and power optimization are jointly utilized for multiuser (MU) recognition in Power domain Non-Orthogonal Multiple Access (PD-NOMA) wireless system. In NOMA systems the successive interference cancellation (SIC) procedure is typically employed at the receiver side, where several users are decoded in a subsequent manner. Fading channels may disperse the transmitted signals and originate dependencies among its samples, this may affect the channel estimation process and consequently affect the SIC process and signal detection accuracy. In this scenario, the impact of Deep Neural Network (DNN) in explicitly estimating the channel coefficients for each user in NOMA cell is investigated. This approach, integrate the Long Short-Term Memory (LSTM) network into the NOMA system where this LSTM network is used for complex data processing to carry out training, updating, and predicting. The DNN is trained online based on channel statistics and then the trained model is used to predict the channel parameters that will be exploited by the receiver in retrieving the original data. Furthermore, Power coefficients are optimized in order to maximize the sum throughput of the system users based on the total transmitted power and Quality of service (QoS) constraints. We formulate an expression for Signal to interference noise ratio (SINR) for each user in the system, then an analysis for the optimization problem and the considered constraints to verify the concavity of the objective function is presented. Lagrange function and Karush–Kuhn–Tucker (KKT) optimality conditions are utilized to derive the optimal power coefficients. Simulation results for different metrics such as bit Error Rate (BER), sum rate, Outage probability and individual user capacity have proved the superiority of the DL approach over the conventional approaches in terms of channel estimation.