Differentiated Embedded Pilot Assisted Automatic Modulation Classification for OTFS System: A Multi-Domain Fusion...

Area of Science:

• Wireless Communications
• Signal Processing
• Machine Learning

Background:

• Orthogonal time-frequency space (OTFS) modulation is crucial for high-mobility environments due to its ability to mitigate Doppler shifts.
• Automatic modulation classification (AMC) is a vital preprocessing step for OTFS systems, yet it remains under-explored.

Purpose of the Study:

• To propose a novel and effective AMC approach for OTFS systems.
• To enhance the accuracy and robustness of modulation classification in challenging wireless conditions.

Main Methods:

• A dual-stream convolutional neural network (CNN) model was developed to extract multi-domain signal features.
• A differentiated embedded pilot structure was designed to improve the discrimination between different modulation schemes.

Main Results:

• The proposed AMC approach achieved high classification accuracy, particularly under low signal-to-noise ratio (SNR) conditions.
• The method demonstrated superior performance compared to existing state-of-the-art baseline approaches.

Conclusions:

• The novel AMC approach offers a significant advancement for OTFS systems, enabling reliable modulation identification.
• The combination of dual-stream CNN and differentiated pilots provides a robust solution for future wireless communication systems.