Optical dispersion effect-based short-time Fourier transform signal recognition method

A photonics-based signal modulation recognition method is proposed. Utilizing optical pulses and dispersion elements, the frequencies of received radio frequency (RF) signals are mapped to time intervals of the output electrical pulses periodically, thereby obtaining time and frequency parameters of RF signals. These parameters are then processed by a convolutional neural network (CNN) to recognize the modulation type. A proof-of-concept experiment demonstrates the proposed method. The verification process confirmed the system's capability to measure frequencies in the range of 1-30 GHz. The model’s robustness is validated by an F1-score of 0.91. And the proposed method achieves a recognition accuracy of 87.09% in classifying six modulation types. The proposed photonics-based parameter measurement technology enables real-time and wideband processing, thereby surpassing digital signal processing limits. The hybrid optical-CNN framework improves accuracy and adaptability, benefiting electronic warfare, radar intelligence, and advanced communications.