OFDM Signal Modulation Classification using Dilated CNN Model
Riddhi Shah, Ashok Parmar, Ankit Chouhan, Kamal M. Captain · 2025
Modern wireless communication systems, including Wi-Fi, LTE, and 5G, use the highly efficient digital transmission technique known as orthogonal frequency division multiplexing (OFDM). Automatic modulation classification (AMC) is a critical signal-processing technique employed in the physical layer of wireless communication systems. Its primary objective is to identify an incoming signal’s modulation type without prior knowledge at the receiver end. This letter outlines our proposed deep learning model for the modulation classification of OFDM using dilated convolution and modified residual network block which uses skip connections for better performance. This model is referred to as a modified residual dilated convolutional neural network (MRDCNN). The dataset containing 6 different modulation techniques which contain different modulation techniques for header and payload of OFDM signals with the Signal-to-Noise Ratio (SNR) range of -10 dB to 20 dB is used here. The implementation results show that the model gives an accuracy of 93.01% for overall SNRs ranging from -10 dB to 20 dB. It can achieve 98% accuracy for 0 dB SNR and can achieve 100% accuracy for SNRs higher than 4 dB. In addition, the proposed model outperforms the state-of-the-art models in modulation classification in OFDM signals.