Resource Allocation and Deep Learning-Based Joint Detection Scheme in Satellite NOMA Systems

Abstract

To overcome the challenges of complex time-varying satellite channels and severe inter-user interference in non-orthogonal multiple access (NOMA), rational power allocation and accurate multi-user joint detection methods are essential. In this paper, a sparrow search algorithm-based resource allocation and deep learning-based joint detection scheme (SSA-DeepJD) in the satellite-terrestrial NOMA system is proposed. First, the NOMA-orthogonal frequency division multiplexing (OFDM) system model is constructed. Next, a convolutional neural network-based image super-resolution recovery network is proposed for offline training and online channel estimation, which incorporates densely connected convolutional layers and residual learning to model for handling complex non-linear channel fitting. Then, a multi-user signal detection based on an iterative deep neural network is proposed, which is iteratively retrained to improve the detection accuracy. Finally, due to the significant impact of the power allocation on the system error performance, the optimal power allocation is found within the power allocation factor threshold based on SSA. Simulation results show that the proposed SSA-DeepJD algorithm is well-suited for multi-user superposed NOMA systems and complex non-linear channel environments. Compared to the baseline algorithms, the SSA-DeepJD algorithm degrades the Bit Error Rate (BER) by 21.5 dB and 11.9 dB in the 2-user and 3-user NOMA systems, respectively.