Photonic-Assisted Modulation Format Identification for RF Signals under Low Sampling Rate

A photonic-assisted approach is proposed for modulation format identification (MFI) on radio frequency (RF) signals under low sampling rate. In this approach, a photonic-assisted interferometer (PAI) is designed for computation-free data augmentation by transforming the signal's phase and frequency variations into modulation format-sensitive amplitude features. A fully connected neural network (FCNN) is used to implement end-to-end MFI. An experiment is conducted on the identification of amplitude-shift keying (ASK), binary phase-shift keying (BPSK), quadrature phase-shift keying (QPSK), frequency-shift keying (FSK), and linear frequency modulation (LFM) signals with signal-to-noise ratios (SNRs) from -10 to 15 dB and carrier frequencies within 5 to 10 GHz. The results show that the proposed photonic-assisted modulation format identifier (PA-MFI) achieves the identification accuracy of 82.44% at 1 GHz sampling rate, which is 10.6% higher than the accuracy of direct modulation format identification (Direct-MFI) without PAI processing.

Automated summary

A photonic-assisted approach for modulation format identification (MFI) on RF signals under low sampling rate is proposed. It utilizes a photonic-assisted interferometer (PAI) for computation-free data augmentation by transforming phase and frequency variations into modulation format-sensitive amplitude features. A fully connected neural network (FCNN) is used for end-to-end MFI implementation. Experimental results show that the proposed photonic-assisted modulation format identifier (PA-MFI) achieves higher identification accuracy compared to direct MFI without PAI processing.