Dual-Component Signal Modulation Classification Based on Deep Convolutional Neural Networks

To address mixed signal modulation classification in complex electronic warfare environments, this paper proposes a dual-component signal modulation classification algorithm using deep learning. The approach employs In-phase and Quadrature (I&Q) modulation for baseband signal decomposition and utilizes a deep convolutional neural network (CNN) with multi-dimensional convolution kernels for feature extraction. By integrating a multi-class loss function, the algorithm achieves precise modulation classification of dual-component signals. Simulation results demonstrate 94.07% classification accuracy at 0 dB Interference-to-Noise Ratio (INR) across four signal types, outperforming existing deep learning-based methods by 9.62%, 11.85%, and 10.23% respectively. Experimental validation confirms the algorithm's effectiveness through close alignment between simulated and practical results.