Kolmogorov–Smirnov test–based modulation classification for NOMA systems

Non-orthogonal multiple access (NOMA) is a key multiple access technique for improving spectrum utilisation in next-generation wireless networks. In NOMA receivers, prior identification of the interfering user’s modulation scheme can substantially reduce signalling overhead and improve successive interference cancellation (SIC). In this paper, a Kolmogorov–Smirnov (KS) test-based automatic modulation classification (AMC) framework is developed for identifying the modulation format of the interfering user in a two-user power-domain NOMA system. The KS test uses differences between empirical distributions to classify modulation without requiring feature extraction or offline training. Simulation results show that the KS-based classifier consistently outperforms the cumulant-based SVM approach, achieving up to an 8% improvement in classification accuracy in the low-SNR region (−10 dB to 0 dB). In addition, the classifier maintains robust performance under moderate channel estimation errors for different observation lengths and NOMA power allocation factors.