A Survey on Automatic Modulation Classification Techniques

Automatic Modulation Classification (AMC) plays a crucial role in intelligent wireless communications receivers, positioned between signal detection and demodulation. It is a significant component in various applications and faces particular challenges when dealing with blind modulation classification. In this context, AMC operates without access to essential information about the transmitted signal and receiver parameters, such as carrier frequency, signal power, timing data, and segment offsets. It also has to contend with the complexities of frequency-selective, multipath fading, and time-varying channels in real-world scenarios. This survey paper provides a detailed analysis of AMC techniques, categorizing them into two classes: traditional and advanced methods. Traditional approaches include Likelihood-Based (LB) and Feature-Based (FB) methods, which have been the foundation of AMC for many years. In recent times, advanced techniques, particularly Deep Learning (DL), have emerged as a transformative force, pushing the boundaries of AMC capabilities. With a strong focus on the core subject, this paper explores the intricacies of AMC strategies employed to classify various modulation schemes, such as Amplitude Shift Keying (ASK), Phase Shift Keying (PSK), Frequency Shift Keying (FSK), Pulse Amplitude Modulation (PAM), and Quadrature Amplitude Modulation (QAM), each with different orders and Signal-to-Noise Ratio (SNR) levels. As the paper concludes, it shifts attention to the future challenges that practitioners and researchers are currently grappling with in the ever-evolving landscape of AMC. We hope that this comprehensive survey is a valuable resource for those involved in the fascinating and essential field of Automatic Modulation Classification.